[{"_1":2,"_378":-5,"_379":-5},"loaderData",{"_3":4,"_91":92,"_346":347},"root",{"_5":6,"_34":35,"_64":65},"ctx",{"_7":8,"_23":-5,"_24":25,"_26":27,"_28":22,"_29":22,"_30":22,"_31":32,"_33":22},"auth",{"_9":10,"_11":-5,"_12":-5,"_13":-5,"_14":-5,"_15":-5,"_16":-5,"_17":-5,"_18":-5,"_19":-5,"_20":-5,"_21":22},"tokenType","session_token","sessionClaims","sessionId","sessionStatus","userId","actor","orgId","orgRole","orgSlug","orgPermissions","factorVerificationAge","isAuthenticated",false,"user","preferencesKeybindings","default","preferencesTheme","light","preferencesFixers","preferencesCopilot","preferencesSelectionAi","notificationCount",0,"flagDenols","ENV",{"_36":37,"_38":39,"_40":41,"_42":43,"_44":45,"_46":47,"_48":49,"_50":51,"_52":53,"_54":55,"_56":57,"_58":59,"_60":61,"_62":63},"REACT_APP_SENTRY_DSN","https://586020d86f05489db63a554690acf1e9@o1398359.ingest.sentry.io/6724499","VAL_RUNTIME_NAME","prod","RENDER_GIT_COMMIT","2f607f7b74b4e52529fb75176d403d8b5b284eb1","RENDER_GIT_BRANCH","main","CLERK_PUBLISHABLE_KEY","pk_live_Y2xlcmsudmFsLnRvd24k","POSTHOG_API_TOKEN","phc_nmoyuAkStEZuUp69rYlg1DZgrIt19Gx0Y4Hwf8tGgJP","POSTHOG_API_HOST","https://wart.val.town","WEB_API_TEMPLATE","https://{val}.web.val.run","APP_URL","https://www.val.town","MODULE_URL","https://esm.town","VANITY_WEB_API_TEMPLATE","https://{label}.val.run/","DENOLS_URL","https://lsp.val.town","NODE_ENV","production","API_URL","https://api.val.town","clerkState",{"_66":67},"__internal_clerk_state",{"_68":69,"_70":45,"_71":72,"_73":72,"_74":22,"_75":72,"_76":72,"_77":72,"_78":72,"_79":80,"_89":22,"_90":22},"__clerk_ssr_state",{"_9":10,"_11":-5,"_12":-5,"_13":-5,"_14":-5,"_15":-5,"_16":-5,"_17":-5,"_18":-5,"_19":-5,"_20":-5,"_21":22},"__publishableKey","__proxyUrl","","__domain","__isSatellite","__signInUrl","__signUpUrl","__afterSignInUrl","__afterSignUpUrl","__clerk_debug",{"_81":22,"_21":22,"_82":72,"_83":84,"_85":72,"_86":45,"_87":22,"_88":72},"isSignedIn","proxyUrl","reason","session-token-and-uat-missing","message","publishableKey","isSatellite","domain","__telemetryDisabled","__telemetryDebug","routes/_app.x.$handle.$name",{"_93":94,"_95":96,"_97":43,"_98":99,"_100":-7,"_101":102,"_150":151,"_324":325,"_326":327,"_335":336,"_300":337,"_340":-5,"_341":-5,"_342":22,"_343":22,"_344":345},"handle","prashamtrivedi","name","resume-parser","branchName","path","prompts/resumePrompts.ts","version","project",{"_103":104,"_95":96,"_105":106,"_107":-5,"_108":22,"_109":110,"_111":72,"_112":-5,"_14":113,"_114":115,"_116":117,"_118":119,"_120":32,"_16":121,"_122":123,"_124":123,"_125":126,"_127":126,"_128":126,"_129":123,"_130":123,"_23":131,"_134":135,"_137":138,"_149":22},"id","7e8f5946-2f13-11f0-8061-569c3dd06744","createdAt",["D",1747042122266],"deletedAt","legacyEnvVars","privacy","unlisted","description","imageUrl","1591349e-125d-4fb7-930d-cf293627961f","pinnedFileId","958b7d88-2f17-11f0-8eca-569c3dd06744","pinPreview",true,"screenshotVersion",39,"likeCount","4d7791fe-1f5b-4867-8b6e-34823c16ee48","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm",{"_93":94,"_132":133},"profileImageUrl","https://img.clerk.com/eyJ0eXBlIjoicHJveHkiLCJzcmMiOiJodHRwczovL2ltYWdlcy5jbGVyay5kZXYvb2F1dGhfZ2l0aHViL2ltZ18yaUF5OXE3VlcwNVRwRU9wSTRmcE1DZ0FSM1UifQ","org",{"_93":94,"_136":133},"avatar","branch",{"_100":139,"_95":43,"_103":140,"_141":104,"_142":-5,"_143":-5,"_144":-5,"_145":-5,"_146":-5,"_147":96,"_148":94},69,"7e8fbf26-2f13-11f0-8061-569c3dd06744","projectId","forkedBranchId","forkedBranchVersion","forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","isOwned","tree",[152,201,217,233,251,268,294,304,313],{"_103":153,"_95":154,"_155":156,"_157":-5,"_158":140,"_159":160,"_161":117,"_105":162,"_100":32,"_163":164,"_165":-5,"_166":22,"_167":-5,"_168":169,"_98":200},"9498b1f2-2f17-11f0-b5a2-569c3dd06744","controllers","fileId","9498b2e2-2f17-11f0-b5a2-569c3dd06744","parentId","branchId","branchVersion",2,"latest",["D",1747043877220],"fileType","directory","content","deleted","label","children",[170,181,190],{"_103":171,"_95":172,"_155":173,"_157":156,"_158":140,"_159":174,"_161":117,"_105":175,"_100":176,"_163":177,"_165":178,"_166":22,"_167":-5,"_168":179,"_98":180},"30d0e38a-2f29-11f0-bc2d-569c3dd06744","healthController.ts","95472a66-2f17-11f0-8cb7-569c3dd06744",33,["D",1747051440754],1,"script","import { OpenAPIHono } from 'npm:@hono/zod-openapi';\nimport { createRoute } from 'npm:@hono/zod-openapi';\nimport { z } from 'npm:@hono/zod-openapi';\n\nconst healthRouter = new OpenAPIHono();\n\n// Define the health response schema\nconst HealthResponseSchema = z.object({\n status: z.string().openapi({\n example: 'ok'\n }),\n version: z.string().openapi({\n example: '1.0.0'\n }),\n timestamp: z.string().openapi({\n example: new Date().toISOString()\n }),\n service: z.string().openapi({\n example: 'resume-parser-api'\n })\n}).openapi('HealthResponse');\n\n// Create the health check route\nconst healthRoute = createRoute({\n method: 'get',\n path: '/health',\n tags: ['System'],\n summary: 'Health Check',\n description: 'Health check endpoint to verify API is operational',\n responses: {\n 200: {\n description: 'API is operational',\n content: {\n 'application/json': {\n schema: HealthResponseSchema\n }\n }\n }\n }\n});\n\n// Register the route\nhealthRouter.openapi(healthRoute, (c) => {\n return c.json({\n status: 'ok',\n version: '1.0.0',\n timestamp: new Date().toISOString(),\n service: 'resume-parser-api'\n });\n});\n\nexport default healthRouter;",[],[154,172],{"_103":182,"_95":183,"_155":184,"_157":156,"_158":140,"_159":185,"_161":117,"_105":186,"_100":32,"_163":177,"_165":187,"_166":22,"_167":-5,"_168":188,"_98":189},"b8da0998-4854-11f0-8366-76b3cceeab13","mcpController.ts","b8da0a42-4854-11f0-8366-76b3cceeab13",59,["D",1749818916417],"\nimport {McpServer} from \"npm:@modelcontextprotocol/sdk/server/mcp.js\"\nimport {StreamableHTTPServerTransport} from \"npm:@modelcontextprotocol/sdk/server/streamableHttp.js\"\n\nimport {OpenAPIHono} from 'npm:@hono/zod-openapi'\nimport {ResumeParserRequestSchema} from \"../schemas/resumeSchemas.ts\"\nimport {extractResumeData} from \"../utils/resumeParser.ts\"\nimport {toFetchResponse, toReqRes} from 'npm:fetch-to-node'\nimport {z} from \"npm:zod\"\n\nconst mcpController = new OpenAPIHono()\n\n\n// Map to store transports by session ID\nconst transports: {[sessionId: string]: StreamableHTTPServerTransport} = {}\n\nmcpController.post(async (c) => {\n const {req, res} = toReqRes(c.req.raw)\n const transport: StreamableHTTPServerTransport = new StreamableHTTPServerTransport({\n sessionIdGenerator: undefined,\n })\n const server = new McpServer({\n name: \"mcp-server\",\n version: \"1.0.0\"\n })\n\n\n server.tool('parse-resume', 'Parse given resume text', {\n resumeText: z.string().describe('Resume text to parse'),\n options: z.object({\n confidenceThreshold: z.number().describe(\n 'Minimum confidence score threshold (0-1)'\n ).default(0.3),\n standardizationEnabled: z.boolean().describe('Whether to standardize terminology').default(true),\n extractSummary: z.boolean().describe('Whether to extract and analyze the resume summary').default(true),\n sectionPriorities: z.array(z.string()).describe('Sections to prioritize in extraction').default([]),\n extractLanguages: z.boolean().describe('Whether to extract language proficiencies').default(true)\n })\n }, async ({resumeText, options}: {resumeText: string, options: any}) => {\n console.error({resume: resumeText, options})\n const parsedResume = await extractResumeData(resumeText, options)\n return {\n isError: false,\n content: [{type: \"text\", text: JSON.stringify(parsedResume)}]\n }\n })\n // Connect to the MCP server\n await server.connect(transport)\n\n\n // Handle the request\n await transport.handleRequest(req, res, await c.req.json())\n res.on('close', () => {\n console.log('Request closed')\n transport.close()\n server.close()\n })\n return toFetchResponse(res)\n})\n\n// mcpController.get(c => {\n// console.log('Received GET MCP request')\n// return c.json({\n// jsonrpc: \"2.0\",\n// error: {\n// code: -32000,\n// message: \"Method not allowed.\"\n// },\n// id: null\n// }, 405)\n// })\n// mcpController.delete(c => {\n// console.log('Received GET MCP request')\n// return c.json({\n// jsonrpc: \"2.0\",\n// error: {\n// code: -32000,\n// message: \"Method not allowed.\"\n// },\n// id: null\n// }, 405)\n// })\n\n\nexport default mcpController",[],[154,183],{"_103":191,"_95":192,"_155":193,"_157":156,"_158":140,"_159":194,"_161":117,"_105":195,"_100":196,"_163":177,"_165":197,"_166":22,"_167":-5,"_168":198,"_98":199},"ed09f24a-2f6e-11f0-a079-569c3dd06744","resumeController.ts","954ac70c-2f17-11f0-b0f4-569c3dd06744",56,["D",1747081391808],4,"import {OpenAPIHono} from 'npm:@hono/zod-openapi'\nimport {createRoute} from 'npm:@hono/zod-openapi'\nimport {extractResumeData} from '../utils/resumeParser.ts'\nimport {\n ResumeParserRequest,\n ParseResumeResponse,\n ParsedResume\n} from '../types/resume.ts'\nimport {\n ResumeParserRequestSchema,\n ParsedResumeSchema,\n ParseResumeResponseSchema,\n ErrorResponseSchema\n} from '../schemas/resumeSchemas.ts'\n\nconst resumeRouter = new OpenAPIHono()\n\n// Define the parse resume route\nconst parseResumeRoute = createRoute({\n method: 'post',\n path: '/parse',\n tags: ['Resume'],\n summary: 'Parse Resume',\n description: 'Parse resume text and extract structured data with confidence scoring',\n request: {\n body: {\n content: {\n 'application/json': {\n schema: ResumeParserRequestSchema\n }\n },\n required: true,\n description: 'Resume text and parsing options'\n }\n },\n responses: {\n 200: {\n description: 'Resume parsed successfully',\n content: {\n 'application/json': {\n schema: ParseResumeResponseSchema\n }\n }\n },\n 400: {\n description: 'Invalid request parameters',\n content: {\n 'application/json': {\n schema: ErrorResponseSchema\n }\n }\n },\n 500: {\n description: 'Server error',\n content: {\n 'application/json': {\n schema: ErrorResponseSchema\n }\n }\n }\n }\n})\n\n// Register the route\nresumeRouter.openapi(parseResumeRoute, async (c) => {\n try {\n // Extract request body with validation\n const {resumeText, options} = c.req.valid('json')\n\n // Extract resume data with confidence scoring\n const parsedResume = await extractResumeData(resumeText, options)\n\n // Return parsed response\n return c.json({\n success: true,\n data: parsedResume\n })\n\n } catch (error) {\n console.error('Resume parsing error:', error)\n\n return c.json({\n success: false,\n error: {\n message: 'Failed to parse resume',\n code: 'PARSING_ERROR',\n details: (error as Error).message\n }\n }, 500)\n }\n})\n\n\n\nexport default resumeRouter",[],[154,192],[154],{"_103":202,"_95":203,"_155":204,"_157":-5,"_158":140,"_159":196,"_161":117,"_105":205,"_100":32,"_163":164,"_165":-5,"_166":22,"_167":-5,"_168":206,"_98":216},"94edccdc-2f17-11f0-a818-569c3dd06744","middleware","94edcd7c-2f17-11f0-a818-569c3dd06744",["D",1747043877778],[207],{"_103":208,"_95":209,"_155":210,"_157":204,"_158":140,"_159":211,"_161":117,"_105":212,"_100":176,"_163":177,"_165":213,"_166":22,"_167":-5,"_168":214,"_98":215},"30d8ff5c-2f29-11f0-8a17-569c3dd06744","errorHandler.ts","957b4a9e-2f17-11f0-8cb7-569c3dd06744",35,["D",1747051440756],"import { Context } from 'npm:hono';\nimport { HTTPException } from 'npm:hono/http-exception';\nimport { ZodError } from 'npm:zod';\n\n// Interface for structured error responses\ninterface ErrorResponse {\n success: false;\n error: {\n message: string;\n code: string;\n details?: unknown;\n };\n}\n\nexport const errorHandler = (err: Error | HTTPException | ZodError, c: Context) => {\n console.error('Error occurred:', err);\n\n let message = 'Internal Server Error';\n let status = 500;\n let code = 'INTERNAL_ERROR';\n let details = undefined;\n\n // Handle HTTPException (Hono's built-in error)\n if (err instanceof HTTPException) {\n message = err.message;\n status = err.status;\n code = `HTTP_${status}`;\n }\n // Handle Zod validation errors\n else if (err instanceof ZodError || err.name === 'ZodError') {\n message = 'Invalid request data';\n status = 400;\n code = 'VALIDATION_ERROR';\n details = err instanceof ZodError ? err.format() : err;\n }\n // Handle validation errors\n else if (err.name === 'ValidationError') {\n message = 'Validation Error';\n status = 400;\n code = 'VALIDATION_ERROR';\n details = err.cause;\n }\n\n // Create structured error response\n const response: ErrorResponse = {\n success: false,\n error: {\n message,\n code,\n ...(details && { details }),\n },\n };\n\n // Log error details\n console.error({\n timestamp: new Date().toISOString(),\n error: err.message,\n stack: err.stack,\n path: c.req.path,\n method: c.req.method,\n status,\n });\n\n return c.json(response, status);\n};",[],[203,209],[203],{"_103":218,"_95":219,"_155":220,"_157":-5,"_158":140,"_159":176,"_161":117,"_105":221,"_100":32,"_163":164,"_165":-5,"_166":22,"_167":-5,"_168":222,"_98":232},"946e3f9e-2f17-11f0-b0f4-569c3dd06744","prompts","946e4066-2f17-11f0-b0f4-569c3dd06744",["D",1747043876941],[223],{"_103":224,"_95":225,"_155":226,"_157":220,"_158":140,"_159":227,"_161":117,"_105":228,"_100":160,"_163":177,"_165":229,"_166":22,"_167":-5,"_168":230,"_98":231},"5ad2f02a-2f64-11f0-8840-569c3dd06744","resumePrompts.ts","954eb786-2f17-11f0-8670-569c3dd06744",53,["D",1747076851535],"/**\n * Generate prompts for different resume parsing operations\n * \n * @param type The type of prompt to generate\n * @returns The prompt text\n */\nexport function generatePrompt(type: 'extract' | 'validate' | 'confidence'): string {\n switch (type) {\n case 'extract':\n return `Extract structured data from the following resume with HIGH PRECISION.\n\n\nIMPORTANT GUIDELINES:\n1. Confidence scores range from 0 to 1, where:\n - 0.9-1.0: Very high confidence (clear, unambiguous data)\n - 0.7-0.9: High confidence (likely correct, but could have minor issues)\n - 0.5-0.7: Medium confidence (possibly correct, but some uncertainty)\n - 0.3-0.5: Low confidence (significant uncertainty)\n - 0.0-0.3: Very low confidence (likely incorrect or missing data)\n\n2. Standardization notes should be provided whenever you transform or normalize the data, such as:\n - Converting abbreviations to full forms (e.g., \"BS\" → \"Bachelor of Science\")\n - Normalizing job titles (e.g., \"Dev\" → \"Developer\")\n - Standardizing technology names (e.g., \"JS\" → \"JavaScript\")\n - Date format standardization\n - If no standardization was performed, use null\n\n3. For missing or unavailable data:\n - Use null for string fields that are missing\n - Use empty array [] for array fields that are missing\n - Use 0 as the confidence score for missing fields\n - Add the field path to missingFields array (e.g., \"personalInfo.website\")\n\n4. For dates:\n - Use ISO format (YYYY-MM) whenever possible\n - Mark current positions with current: true and null endDate\n - Calculate durationInMonths when possible\n\n5. For skills:\n - Infer categories when possible (e.g., \"Programming Languages\", \"Tools\", \"Frameworks\")\n - Infer proficiency when mentioned (\"Beginner\", \"Intermediate\", \"Advanced\", \"Expert\")\n - Detect years of experience when mentioned\n\n6. For work experience:\n - Extract technologies used in each role when mentioned\n - Separate responsibilities from achievements when possible\n - Standardize job titles to industry norms\n\n7. DO NOT fabricate or assume information not present in the resume. If information is missing, use null with a confidence score of 0.\n\n8. The overallConfidence score should be a weighted average of all fields, giving more weight to critical fields like name, contact info, work experience, and education.\n\n9. The detectedSections array should list all major sections found in the resume (e.g., personal_information, skills, experience, education, etc.).`;\n \n case 'validate':\n return `Validate the following resume data for consistency, accuracy, and standardization.\n\nFor each field in the data:\n1. Check for inconsistencies and validate format\n2. Standardize terminology where appropriate:\n - Degree names (e.g., \"BS\" → \"Bachelor of Science\", \"MS\" → \"Master of Science\")\n - Job titles (e.g., \"SWE\" → \"Software Engineer\", \"PM\" → \"Product Manager\")\n - Technology names (e.g., \"JS\" → \"JavaScript\", \"TS\" → \"TypeScript\")\n - Institution names (e.g., \"NYU\" → \"New York University\")\n3. Validate dates:\n - Ensure date ranges are logical (start before end)\n - Verify that duration calculations are accurate\n - Check for overlapping or inconsistent date ranges\n4. Look for contradictions:\n - Skills mentioned in work experience but not in skills section\n - Technologies that don't match the responsibilities\n - Date inconsistencies between sections\n5. Update confidence scores:\n - Increase confidence for validated fields\n - Decrease confidence for fields with detected issues\n - Add standardization notes when changes are made\n\nReturn the validated and standardized data in the exact same format, including:\n- All confidence scores updated based on validation\n- Standardization notes added for any changes made\n- Overall confidence score recalculated\n- Missing fields list updated if additional missing fields are found\n\nIf you detect serious inconsistencies or potential fabrications, mark those fields with lower confidence scores and add appropriate standardization notes.`;\n \n case 'confidence':\n return `Analyze the following resume data and calculate confidence scores for each field.\n\nConfidence scoring criteria by field type:\n\n1. Personal Information:\n - Name: Check for full name vs. partial, common formats\n - Email: Validate format, check for standard patterns\n - Phone: Validate format, check country code presence\n - Location: Check for specificity (city/state/country)\n - Social/Web: Validate URL formats and completeness\n\n2. Skills:\n - Compare against known technology/skill names\n - Check for specificity (general vs. specific skills)\n - Look for proficiency indicators\n - Consider consistency with work experience\n\n3. Work Experience:\n - Validate company names against known entities\n - Check job title standardization\n - Verify date formats and calculations\n - Evaluate responsibility descriptions (specific vs. vague)\n - Look for quantifiable achievements\n - Check for technology mentions that match skills\n\n4. Education:\n - Validate institution names\n - Check degree nomenclature standardization\n - Verify field of study specificity\n - Validate date formats and ranges\n - Check GPA format and range\n - Evaluate coursework relevance\n\n5. Projects & Certifications:\n - Evaluate description specificity\n - Check URL validity\n - Verify technology consistency with skills\n - Validate certification issuers\n - Check date formats and validity\n\nRate each field on a scale from 0 to 1, where:\n- 0.9-1.0: Very high confidence (clear, unambiguous data)\n- 0.7-0.9: High confidence (likely correct, but could have minor issues)\n- 0.5-0.7: Medium confidence (possibly correct, but some uncertainty)\n- 0.3-0.5: Low confidence (significant uncertainty)\n- 0.0-0.3: Very low confidence (likely incorrect or missing data)\n\nFor any field where confidence is below 0.7, include specific reasons in the standardization notes.\n\nReturn the data with confidence scores added to each field and an overall confidence score that is a weighted average of all fields.`;\n \n default:\n return '';\n }\n}",[],[219,225],[219],{"_103":234,"_95":235,"_155":236,"_157":-5,"_158":140,"_159":237,"_161":117,"_105":238,"_100":32,"_163":164,"_165":-5,"_166":22,"_167":-5,"_168":239,"_98":250},"e613cb0a-2f28-11f0-b8b3-569c3dd06744","schemas","e613cbdc-2f28-11f0-b8b3-569c3dd06744",30,["D",1747051315367],[240],{"_103":241,"_95":242,"_155":243,"_157":236,"_158":140,"_159":244,"_161":117,"_105":245,"_100":246,"_163":177,"_165":247,"_166":22,"_167":-5,"_168":248,"_98":249},"68f5fa44-2f64-11f0-948a-569c3dd06744","resumeSchemas.ts","e63ea708-2f28-11f0-a42c-569c3dd06744",55,["D",1747076875254],9,"import {z} from 'npm:@hono/zod-openapi'\nimport {ZodObject, ZodType, ZodTypeAny} from \"npm:zod\"\n\n/**\n * Transforms a Zod schema into a format suitable for OpenAI by:\n * - Adding additionalProperties: false to all objects\n * - Making all fields required\n * - Removing validation constraints that OpenAI might struggle with\n *\n * Specifically removes:\n * - For strings: minLength, maxLength, pattern, format (min, max, regex, email, url, etc.)\n * - For numbers: minimum, maximum, multipleOf (min, max, multipleOf, int, etc.)\n * - For objects: patternProperties, unevaluatedProperties, propertyNames, etc.\n * - For arrays: minItems, maxItems, uniqueItems (min, max, length, nonempty, etc.)\n *\n * @param schema The original Zod schema to transform\n * @returns A new Zod schema with modified properties\n */\nexport function transformSchemaForOpenAI(schema: T): T {\n // Helper to process each schema type\n function processSchema(schema: ZodTypeAny): ZodTypeAny {\n // Handle different schema types\n if (schema instanceof z.ZodObject) {\n const shape = schema._def.shape()\n const newShape: Record = {}\n\n // Process each field in the object\n for (const [key, value] of Object.entries(shape)) {\n newShape[key] = processSchema(value)\n }\n\n // Create new object schema with additionalProperties: false and all fields required\n return z.object(newShape).openapi({\n // Preserve description and example if they exist\n ...(schema._def.openapi ? {\n description: schema._def.openapi.description,\n example: schema._def.openapi.example\n } : {}),\n // Always add additionalProperties: false\n additionalProperties: false,\n })\n }\n else if (schema instanceof z.ZodArray) {\n // Process array items, removing all array constraints\n // This removes: minItems, maxItems, uniqueItems validation\n return z.array(processSchema(schema._def.type)).openapi({\n // Preserve description and example if they exist\n ...(schema._def.openapi ? {\n description: schema._def.openapi.description,\n example: schema._def.openapi.example\n } : {})\n })\n }\n else if (schema instanceof z.ZodString) {\n // Remove ALL string validations (min, max, regex, email, url, etc.)\n return z.string().openapi({\n // Preserve description and example if they exist\n ...(schema._def.openapi ? {\n description: schema._def.openapi.description,\n example: schema._def.openapi.example\n } : {})\n })\n }\n else if (schema instanceof z.ZodNumber) {\n // Remove ALL number validations (min, max, multipleOf, int, etc.)\n return z.number().openapi({\n // Preserve description and example if they exist\n ...(schema._def.openapi ? {\n description: schema._def.openapi.description,\n example: schema._def.openapi.example\n } : {})\n })\n }\n else if (schema instanceof z.ZodBoolean) {\n // Preserve boolean as is\n return z.boolean().openapi({\n // Preserve description and example if they exist\n ...(schema._def.openapi ? {\n description: schema._def.openapi.description,\n example: schema._def.openapi.example\n } : {})\n })\n }\n else if (schema instanceof z.ZodNullable) {\n // Handle nullable fields\n return z.nullable(processSchema(schema.unwrap()))\n }\n else if (schema instanceof z.ZodOptional) {\n // Make optional fields required\n return processSchema(schema.unwrap())\n }\n else if (schema instanceof z.ZodEnum || schema instanceof z.ZodLiteral) {\n // Preserve enums and literals but without additional constraints\n // Convert them to strings for maximum compatibility\n return z.string().openapi({\n // Preserve description and example if they exist\n ...(schema._def.openapi ? {\n description: schema._def.openapi.description,\n example: schema._def.openapi.example\n } : {})\n })\n }\n else if (schema instanceof z.ZodUnion || schema instanceof z.ZodDiscriminatedUnion) {\n // For unions, just convert to the first type for simplicity\n // This is a simplification that might lose some information\n // but ensures OpenAI has a concrete type to work with\n return processSchema(schema._def.options[0])\n }\n\n // Return other schema types unchanged\n return schema\n }\n\n return processSchema(schema) as T\n}\n\n// Generic confidence field schema\nconst ConfidenceFieldSchema = (valueSchema: T) =>\n z.object({\n value: valueSchema,\n confidence: z.number().min(0).max(1).openapi({\n description: 'Confidence score between 0 and 1',\n example: 0.95\n }),\n standardization: z.string().nullable().openapi({\n description: 'Notes about any standardization applied',\n example: 'Standardized from \"JS\" to \"JavaScript\"'\n })\n })\n\n// Date range schema\nconst DateRangeSchema = z.object({\n startDate: ConfidenceFieldSchema(z.string().nullable()),\n endDate: ConfidenceFieldSchema(z.string().nullable()),\n durationInMonths: ConfidenceFieldSchema(z.number().nullable()),\n current: ConfidenceFieldSchema(z.boolean())\n}).openapi('DateRange')\n\n// Request schema\nexport const ResumeParserRequestSchema = z.object({\n resumeText: z.string().min(1).openapi({\n description: 'The raw text of the resume to parse',\n example: 'John Doe\\nSoftware Engineer\\n...'\n }),\n options: z.object({\n confidenceThreshold: z.number().min(0).max(1).optional().openapi({\n description: 'Minimum confidence score threshold (0-1)',\n example: 0.5\n }),\n standardizationEnabled: z.boolean().optional().openapi({\n description: 'Whether to standardize terminology',\n example: true\n }),\n extractSummary: z.boolean().optional().openapi({\n description: 'Whether to extract and analyze the resume summary',\n example: true\n }),\n sectionPriorities: z.array(z.string()).optional().openapi({\n description: 'Sections to prioritize in extraction',\n example: ['skills', 'workExperience']\n }),\n extractLanguages: z.boolean().optional().openapi({\n description: 'Whether to extract language proficiencies',\n example: true\n })\n }).optional().openapi({\n description: 'Parser configuration options'\n })\n}).openapi('ResumeParserRequest')\n\n// Personal info schema\nconst PersonalInfoSchema = z.object({\n name: ConfidenceFieldSchema(z.string()),\n email: ConfidenceFieldSchema(z.string()),\n phone: ConfidenceFieldSchema(z.string()),\n location: ConfidenceFieldSchema(z.string().nullable()),\n linkedin: ConfidenceFieldSchema(z.string().nullable()),\n github: ConfidenceFieldSchema(z.string().nullable()),\n website: ConfidenceFieldSchema(z.string().nullable()),\n summary: ConfidenceFieldSchema(z.string().nullable())\n}).openapi('PersonalInfo')\n\n// Skill schema\nconst SkillSchema = z.object({\n name: ConfidenceFieldSchema(z.string()),\n category: ConfidenceFieldSchema(z.string().nullable()),\n proficiency: ConfidenceFieldSchema(z.string().nullable()),\n yearsOfExperience: ConfidenceFieldSchema(z.number().nullable())\n}).openapi('Skill')\n\n// Work experience schema\nconst WorkExperienceSchema = z.object({\n company: ConfidenceFieldSchema(z.string()),\n title: ConfidenceFieldSchema(z.string()),\n location: ConfidenceFieldSchema(z.string().nullable()),\n dates: DateRangeSchema,\n responsibilities: ConfidenceFieldSchema(z.array(z.string())),\n technologies: ConfidenceFieldSchema(z.array(z.string())),\n achievements: ConfidenceFieldSchema(z.array(z.string()))\n}).openapi('WorkExperience')\n\n// Education schema\nconst EducationSchema = z.object({\n institution: ConfidenceFieldSchema(z.string()),\n degree: ConfidenceFieldSchema(z.string()),\n field: ConfidenceFieldSchema(z.string()),\n dates: DateRangeSchema,\n gpa: ConfidenceFieldSchema(z.string().nullable()),\n coursework: ConfidenceFieldSchema(z.array(z.string()).nullable()),\n achievements: ConfidenceFieldSchema(z.array(z.string()).nullable())\n}).openapi('Education')\n\n// Project schema\nconst ProjectSchema = z.object({\n name: ConfidenceFieldSchema(z.string()),\n description: ConfidenceFieldSchema(z.string()),\n technologies: ConfidenceFieldSchema(z.array(z.string())),\n urls: ConfidenceFieldSchema(z.array(z.string()).nullable()),\n dates: ConfidenceFieldSchema(DateRangeSchema.nullable()),\n role: ConfidenceFieldSchema(z.string().nullable())\n}).openapi('Project')\n\n// Certification schema\nconst CertificationSchema = z.object({\n name: ConfidenceFieldSchema(z.string()),\n issuer: ConfidenceFieldSchema(z.string().nullable()),\n date: ConfidenceFieldSchema(z.string().nullable()),\n expiryDate: ConfidenceFieldSchema(z.string().nullable()),\n id: ConfidenceFieldSchema(z.string().nullable())\n}).openapi('Certification')\n\n// Language schema\nconst LanguageSchema = z.object({\n name: ConfidenceFieldSchema(z.string()),\n proficiency: ConfidenceFieldSchema(z.string().nullable())\n}).openapi('Language')\n\n// Parsed resume schema\nexport const ParsedResumeSchema = z.object({\n personalInfo: PersonalInfoSchema,\n skills: z.array(SkillSchema),\n workExperience: z.array(WorkExperienceSchema),\n education: z.array(EducationSchema),\n projects: z.array(ProjectSchema),\n certifications: z.array(CertificationSchema),\n languages: z.array(LanguageSchema),\n overallConfidence: z.number().min(0).max(1),\n missingFields: z.array(z.string()),\n detectedSections: z.array(z.string())\n}).openapi('ParsedResume')\n\n// Response schema\nexport const ParseResumeResponseSchema = z.object({\n success: z.boolean(),\n data: ParsedResumeSchema.optional(),\n error: z.object({\n message: z.string(),\n code: z.string(),\n details: z.unknown().optional()\n }).optional()\n}).openapi('ParseResumeResponse')\n\n// Error response schema\nexport const ErrorResponseSchema = z.object({\n success: z.literal(false),\n error: z.object({\n message: z.string(),\n code: z.string(),\n details: z.unknown().optional()\n })\n}).openapi('ErrorResponse')\n//Simplified ConfidenceFieldSchema\nexport const SimplifiedConfidenceFieldSchema = (valueSchema: T) =>\n z.object({\n value: valueSchema.describe(\"The original value passed, keep this as it is.\"),\n confidence: z.number().describe(\"A confidence score between 0 to 1 about clarity and standardise termwords, 0 is for not clear and not standard and 1 is for clear and standard terms. Abbreviations are not standard\"),\n standardization: z.string().describe(\"If the value has confidence > 0.8 keep pass the original value as is, Otherwise pass in following format, suitable for tooltip: UpdatedStandardValue=ReasonBehindUpdation\")\n })\n// Create a simplified schema for OpenAI with fewer parameters\n// This addresses the 100 parameter limit in OpenAI's response_format\nexport const SimplifiedResumeSchema = z.object({\n personalInfo: z.object({\n name: z.string(),\n email: z.string(),\n phone: z.string(),\n location: z.string().nullable(),\n linkedin: z.string().nullable(),\n github: z.string().nullable(),\n website: z.string().nullable(),\n summary: z.string().nullable()\n }),\n skills: z.array(z.object({\n name: SimplifiedConfidenceFieldSchema(z.string()),\n category: SimplifiedConfidenceFieldSchema(z.string().nullable()),\n proficiency: SimplifiedConfidenceFieldSchema(z.string().nullable()),\n yearsOfExperience: z.number().nullable()\n })),\n workExperience: z.array(z.object({\n company: z.string(),\n title: z.string(),\n location: z.string().nullable(),\n dates: z.object({\n startDate: z.string().nullable(),\n endDate: z.string().nullable(),\n current: z.boolean()\n }),\n responsibilities: z.array(z.string()),\n technologies: z.array(z.string()),\n achievements: z.array(z.string())\n })),\n education: z.array(z.object({\n institution: z.string(),\n degree: SimplifiedConfidenceFieldSchema(z.string()),\n field: SimplifiedConfidenceFieldSchema(z.string()),\n dates: z.object({\n startDate: z.string().nullable(),\n endDate: z.string().nullable(),\n current: z.boolean()\n }),\n gpa: z.string().nullable(),\n coursework: z.array(z.string()).nullable(),\n achievements: z.array(z.string()).nullable()\n })),\n projects: z.array(z.object({\n name: z.string(),\n description: z.string(),\n technologies: SimplifiedConfidenceFieldSchema((z.array(z.string()))),\n urls: z.array(z.string()).nullable(),\n role: SimplifiedConfidenceFieldSchema(z.string().nullable())\n })),\n certifications: z.array(z.object({\n name: SimplifiedConfidenceFieldSchema(z.string()),\n issuer: SimplifiedConfidenceFieldSchema(z.string().nullable()),\n date: z.string().nullable(),\n expiryDate: z.string().nullable(),\n id: z.string().nullable()\n })),\n languages: z.array(z.object({\n name: SimplifiedConfidenceFieldSchema(z.string()),\n proficiency: SimplifiedConfidenceFieldSchema(z.string().nullable())\n })),\n overallConfidence: z.number(),\n missingFields: z.array(z.string()),\n detectedSections: z.array(z.string())\n})\n\n// Example of how to use the transformSchemaForOpenAI function:\n// This creates a version of SimplifiedResumeSchema that's suitable for OpenAI\nexport const OpenAICompatibleResumeSchema = transformSchemaForOpenAI(SimplifiedResumeSchema)\n",[],[235,242],[235],{"_103":252,"_95":253,"_155":254,"_157":-5,"_158":140,"_159":255,"_161":117,"_105":256,"_100":32,"_163":164,"_165":-5,"_166":22,"_167":-5,"_168":257,"_98":267},"9519051e-2f17-11f0-8409-569c3dd06744","types","951905e6-2f17-11f0-8409-569c3dd06744",5,["D",1747043878062],[258],{"_103":259,"_95":260,"_155":261,"_157":254,"_158":140,"_159":262,"_161":117,"_105":263,"_100":176,"_163":177,"_165":264,"_166":22,"_167":-5,"_168":265,"_98":266},"451e0cc2-2f25-11f0-8bdc-569c3dd06744","resume.ts","95b504dc-2f17-11f0-ad48-569c3dd06744",22,["D",1747049756828],"// Generic confidence field type for all extractable fields\nexport interface ConfidenceField {\n value: T;\n confidence: number;\n standardization: string | null;\n}\n\n// Request types\nexport interface ResumeParserRequest {\n resumeText: string;\n options?: ResumeParserOptions;\n}\n\nexport interface ResumeParserOptions {\n confidenceThreshold?: number;\n standardizationEnabled?: boolean;\n extractSummary?: boolean;\n sectionPriorities?: string[];\n extractLanguages?: boolean;\n}\n\n// Enhanced schema with confidence scoring\nexport interface PersonalInfo {\n name: ConfidenceField;\n email: ConfidenceField;\n phone: ConfidenceField;\n location: ConfidenceField;\n linkedin: ConfidenceField;\n github: ConfidenceField;\n website: ConfidenceField;\n summary: ConfidenceField;\n}\n\nexport interface Skill {\n name: ConfidenceField;\n category: ConfidenceField;\n proficiency: ConfidenceField;\n yearsOfExperience: ConfidenceField;\n}\n\nexport interface DateRange {\n startDate: ConfidenceField;\n endDate: ConfidenceField;\n durationInMonths: ConfidenceField;\n current: ConfidenceField;\n}\n\nexport interface WorkExperience {\n company: ConfidenceField;\n title: ConfidenceField;\n location: ConfidenceField;\n dates: DateRange;\n responsibilities: ConfidenceField;\n technologies: ConfidenceField;\n achievements: ConfidenceField;\n}\n\nexport interface Education {\n institution: ConfidenceField;\n degree: ConfidenceField;\n field: ConfidenceField;\n dates: DateRange;\n gpa: ConfidenceField;\n coursework: ConfidenceField;\n achievements: ConfidenceField;\n}\n\nexport interface Project {\n name: ConfidenceField;\n description: ConfidenceField;\n technologies: ConfidenceField;\n urls: ConfidenceField;\n dates: ConfidenceField;\n role: ConfidenceField;\n}\n\nexport interface Certification {\n name: ConfidenceField;\n issuer: ConfidenceField;\n date: ConfidenceField;\n expiryDate: ConfidenceField;\n id: ConfidenceField;\n}\n\nexport interface Language {\n name: ConfidenceField;\n proficiency: ConfidenceField;\n}\n\n// Main resume data structure\nexport interface ParsedResume {\n personalInfo: PersonalInfo;\n skills: Skill[];\n workExperience: WorkExperience[];\n education: Education[];\n projects: Project[];\n certifications: Certification[];\n languages: Language[];\n overallConfidence: number;\n missingFields: string[];\n detectedSections: string[];\n}\n\n// Response interface\nexport interface ParseResumeResponse {\n success: boolean;\n data?: ParsedResume;\n error?: {\n message: string;\n code: string;\n details?: unknown;\n };\n}\n\n// For backward compatibility with existing code\nexport type ContactDetails = PersonalInfo;\nexport type ResumeData = ParsedResume;",[],[253,260],[253],{"_103":269,"_95":270,"_155":271,"_157":-5,"_158":140,"_159":272,"_161":117,"_105":273,"_100":32,"_163":164,"_165":-5,"_166":22,"_167":-5,"_168":274,"_98":293},"94c38b02-2f17-11f0-9ec4-569c3dd06744","utils","94c38bc0-2f17-11f0-9ec4-569c3dd06744",3,["D",1747043877502],[275,284],{"_103":276,"_95":277,"_155":278,"_157":271,"_158":140,"_159":139,"_161":117,"_105":279,"_100":280,"_163":177,"_165":281,"_166":22,"_167":-5,"_168":282,"_98":283},"4b7d3b54-4859-11f0-a120-76b3cceeab13","openai.ts","9567ad72-2f17-11f0-bf12-569c3dd06744",["D",1749820880414],17,"import {OpenAI} from \"npm:openai@5.3.0\"\nimport {ParsedResume, ResumeParserOptions, ConfidenceField, DateRange} from \"../types/resume.ts\"\nimport {generatePrompt} from \"../prompts/resumePrompts.ts\"\nimport {OpenAICompatibleResumeSchema, ParsedResumeSchema} from \"../schemas/resumeSchemas.ts\"\nimport {z} from \"npm:@hono/zod-openapi\"\nimport {zodResponseFormat} from \"npm:openai@5.3.0/helpers/zod\"\n\n/**\n * Parse resume text using OpenAI and return structured data\n *\n * @param resumeText The plain text of the resume to parse\n * @param options Optional parsing configuration\n * @returns Structured resume data with confidence scores\n */\nexport async function parseResumeWithAI(\n resumeText: string,\n options?: ResumeParserOptions\n): Promise {\n try {\n // Input validation\n if (!resumeText || resumeText.trim().length < 50) {\n throw new Error(\"Resume text is too short or empty\")\n }\n\n // Initialize OpenAI client (uses API key from environment in ValTown)\n const openai = new OpenAI({\n apiKey: Deno.env.get(\"OPENAI_API_KEY\")\n })\n\n // Get extraction prompt\n const extractPrompt = generatePrompt('extract')\n\n // Configure extraction based on options\n const extractionFocus = options?.sectionPriorities?.length\n ? `\\n\\nPay special attention to extracting these sections with high accuracy: ${options.sectionPriorities.join(', ')}.`\n : ''\n\n const extractLanguages = options?.extractLanguages === false\n ? '\\n\\nSkip extracting language proficiencies as they are not required.'\n : ''\n\n const extractSummary = options?.extractSummary === false\n ? '\\n\\nSkip extracting and analyzing the resume summary as it is not required.'\n : '\\n\\nEnsure you extract and analyze the resume summary to understand the candidate\\'s self-presentation.'\n\n // Create enhanced system prompt with specific instructions\n const systemPrompt = `You are an expert resume parsing AI specifically trained to extract structured data from resumes with high precision.\n\nYour goal is to analyze the resume text and extract information following these principles:\n1. Extract ALL relevant information into the specified JSON structure\n2. Assign appropriate confidence scores to each field (0-1)\n3. Provide standardization notes when you normalize or transform data\n4. NEVER invent or hallucinate data not present in the resume\n5. For missing data, use null values with confidence scores of 0\n6. Categorize skills appropriately and identify proficiency levels when present\n7. For work experience, separate responsibilities from achievements\n8. Standardize job titles and technology names to common industry terms\n9. Detect and extract project information with associated technologies\n10. Look for education details with specific degree names and institutions\n\n${extractionFocus}${extractLanguages}${extractSummary}`\n\n console.log(\"Calling OpenAI API for initial extraction...\")\n\n console.log({isOpenAI: !!openai.chat.completions})\n\n\n // Call OpenAI API with enhanced system prompt and structured output schema\n const completion = await openai.chat.completions.parse({\n model: \"gpt-4o\",\n messages: [\n {role: \"system\", content: systemPrompt},\n {role: \"user\", content: extractPrompt + \"\\n\\nResume Text:\\n\" + resumeText}\n ],\n response_format: zodResponseFormat(OpenAICompatibleResumeSchema, 'parsedResume')\n })\n try {\n\n // Parse and validate the response\n const responseContent = completion.choices[0]?.message?.parsed\n\n if (!responseContent) {\n throw new Error(\"Empty response from OpenAI\")\n }\n\n // Convert simplified schema format to full ParsedResume format\n return convertToParsedResume(responseContent)\n } catch (error) {\n\n console.error(\"JSON parsing error:\", error)\n throw new Error(\"Failed to parse OpenAI response as valid JSON\")\n }\n } catch (error) {\n console.error(\"OpenAI parsing error:\", error)\n // Provide specific error message based on the error type\n if ((error as Error).message.includes(\"429\")) {\n throw new Error(\"OpenAI rate limit exceeded. Please try again later.\")\n } else if ((error as Error).message.includes(\"401\") || (error as Error).message.includes(\"403\")) {\n throw new Error(\"Authentication error with OpenAI API. Check your API key.\")\n } else {\n throw new Error(\"Failed to parse resume with OpenAI: \" + (error as Error).message)\n }\n }\n}\n\n/**\n * Validate extracted data using OpenAI\n *\n * @param data The extracted resume data to validate\n * @returns Validated resume data with confidence scores\n */\nexport async function validateResumeWithAI(data: ParsedResume): Promise {\n try {\n // Initialize OpenAI client\n const openai = new OpenAI()\n\n // Get validation prompt\n const validationPrompt = generatePrompt('validate')\n\n // Enhanced system prompt for validation\n const systemPrompt = `You are an expert resume data validator specialized in standardizing and improving extracted resume data.\n\nYour task is to analyze the provided JSON resume data and:\n1. Validate all fields for correctness and consistency\n2. Standardize terminology such as:\n - Convert degree abbreviations to full forms (e.g., \"BS\" → \"Bachelor of Science\")\n - Normalize job titles to industry standard terms (e.g., \"Dev\" → \"Developer\", \"SWE\" → \"Software Engineer\")\n - Standardize technology names (e.g., \"JS\" → \"JavaScript\", \"React.js\" → \"React\")\n - Format dates consistently in YYYY-MM format when possible\n3. Cross-reference information for inconsistencies:\n - Skills mentioned in experience but missing from skills section\n - Date overlaps or impossible timelines\n - Educational degrees that don't align with stated institutions\n4. Update confidence scores based on your validation:\n - Increase scores for validated fields\n - Decrease scores for fields with detected issues\n5. Add standardization notes for any changes you make\n\nDO NOT remove or invent data - only standardize, correct, and validate what is provided.\nAlways return the complete JSON structure with your improvements.`\n\n console.log(\"Calling OpenAI API for validation and standardization...\")\n\n\n // Call OpenAI API with enhanced validation\n const completion = await openai.chat.completions.parse({\n model: \"gpt-4o\",\n messages: [\n {role: \"system\", content: systemPrompt},\n {role: \"user\", content: validationPrompt + \"\\n\\nResume Data:\\n\" + JSON.stringify(data)}\n ],\n response_format: zodResponseFormat(OpenAICompatibleResumeSchema, 'validResume')\n })\n\n // Parse and validate the response\n try {\n const responseContent = completion.choices[0]?.message?.parsed\n\n if (!responseContent) {\n throw new Error(\"Empty validation response from OpenAI\")\n }\n\n // Convert simplified schema format to full ParsedResume format\n return convertToParsedResume(responseContent)\n } catch (error) {\n console.error(\"Validation JSON parsing error:\", error)\n console.warn(\"Using original data due to validation error\")\n return data // Return original data if validation fails\n }\n } catch (error) {\n console.error(\"OpenAI validation error:\", error)\n console.warn(\"Using original data due to validation API error\")\n // Return original data if validation API call fails\n return data\n }\n}\n\n/**\n * Helper function to check if OpenAI extraction result is valid\n * and retry if necessary with a different prompt strategy\n *\n * @param data Initial extraction result\n * @param resumeText Original resume text\n * @returns Fixed data or the original data\n */\nexport async function retryExtractionIfNeeded(data: ParsedResume, resumeText: string): Promise {\n try {\n // Check for common extraction issues\n const hasEmptyRequiredFields = !data.personalInfo?.name?.value ||\n !data.personalInfo?.email?.value ||\n data.skills.length === 0 ||\n data.workExperience.length === 0\n\n const hasLowConfidence = data.overallConfidence < 0.5\n\n // If no major issues, return the data as is\n if (!hasEmptyRequiredFields && !hasLowConfidence) {\n console.log(\"All good! Returning the data for next step\")\n return data\n }\n\n console.log(\"Initial extraction had issues, attempting retry with different strategy...\")\n\n // Initialize OpenAI client\n const openai = new OpenAI()\n\n // Create a simplified extraction prompt focused on missing data\n const systemPrompt = `You are an expert resume parser. The initial extraction of this resume had issues.\nPlease focus on extracting these critical elements:\n\n1. Personal information (especially name, email, phone)\n2. Skills and technologies with appropriate categories\n3. Work experience with clear company names, titles, and dates\n4. Education details with institutions and degrees\n\nMaintain the exact same JSON format as the original extraction.\nDo NOT invent information not present in the resume.`\n\n\n // Call OpenAI API with retry strategy\n const completion = await openai.chat.completions.parse({\n model: \"gpt-4o\",\n messages: [\n {role: \"system\", content: systemPrompt},\n {\n role: \"user\", content: `Here is the resume text. Extract the missing information and produce a complete JSON result:\n\nResume:\n${resumeText}\n\nCurrent extraction (with issues):\n${JSON.stringify(data, null, 2)}\n\nPlease provide a complete and corrected JSON response following the exact same schema.` }\n ],\n response_format: zodResponseFormat(OpenAICompatibleResumeSchema, 'retriedResume')\n })\n\n // Parse and return the response\n try {\n const responseContent = completion.choices[0]?.message?.parsed\n\n // Convert simplified schema format to full ParsedResume format\n return convertToParsedResume(responseContent)\n } catch (error) {\n if (error instanceof z.ZodError) {\n console.error(\"Retry Zod validation error:\", error.format())\n return data\n }\n console.error(\"Retry JSON parsing error:\", error)\n return data // Return original data if parsing fails\n }\n } catch (error) {\n console.error(\"Retry extraction error:\", error)\n return data // Return original data if retry fails\n }\n}\n\n/**\n * Convert simplified OpenAI schema format to full ParsedResume format with confidence fields\n * @param simplified Simplified resume data returned from OpenAI\n * @returns Fully structured resume data with confidence fields\n */\nfunction convertToParsedResume(simplified: any): ParsedResume {\n if (!simplified) return simplified\n\n // Helper to create confidence field\n function createConfidenceField(value: T, confidence = 0.5): ConfidenceField {\n return {\n value,\n confidence,\n standardization: null\n }\n }\n\n // Helper to create date range with confidence fields\n function createDateRangeWithConfidence(dates: any): DateRange {\n if (!dates) {\n return {\n startDate: createConfidenceField(null),\n endDate: createConfidenceField(null),\n durationInMonths: createConfidenceField(null),\n current: createConfidenceField(false)\n }\n }\n\n return {\n startDate: createConfidenceField(dates.startDate),\n endDate: createConfidenceField(dates.endDate),\n durationInMonths: createConfidenceField(null), // Will be calculated in post-processing\n current: createConfidenceField(dates.current || false)\n }\n }\n\n // Create full ParsedResume structure with confidence fields\n const result: ParsedResume = {\n personalInfo: {\n name: createConfidenceField(simplified.personalInfo.name),\n email: createConfidenceField(simplified.personalInfo.email),\n phone: createConfidenceField(simplified.personalInfo.phone),\n location: createConfidenceField(simplified.personalInfo.location),\n linkedin: createConfidenceField(simplified.personalInfo.linkedin),\n github: createConfidenceField(simplified.personalInfo.github),\n website: createConfidenceField(simplified.personalInfo.website),\n summary: createConfidenceField(simplified.personalInfo.summary)\n },\n skills: (simplified.skills || []).map((skill: any) => ({\n name: createConfidenceField(skill.name),\n category: createConfidenceField(skill.category),\n proficiency: createConfidenceField(skill.proficiency),\n yearsOfExperience: createConfidenceField(skill.yearsOfExperience)\n })),\n workExperience: (simplified.workExperience || []).map((exp: any) => ({\n company: createConfidenceField(exp.company),\n title: createConfidenceField(exp.title),\n location: createConfidenceField(exp.location),\n dates: createDateRangeWithConfidence(exp.dates),\n responsibilities: createConfidenceField(exp.responsibilities || []),\n technologies: createConfidenceField(exp.technologies || []),\n achievements: createConfidenceField(exp.achievements || [])\n })),\n education: (simplified.education || []).map((edu: any) => ({\n institution: createConfidenceField(edu.institution),\n degree: createConfidenceField(edu.degree),\n field: createConfidenceField(edu.field),\n dates: createDateRangeWithConfidence(edu.dates),\n gpa: createConfidenceField(edu.gpa),\n coursework: createConfidenceField(edu.coursework),\n achievements: createConfidenceField(edu.achievements)\n })),\n projects: (simplified.projects || []).map((proj: any) => ({\n name: createConfidenceField(proj.name),\n description: createConfidenceField(proj.description),\n technologies: createConfidenceField(proj.technologies || []),\n urls: createConfidenceField(proj.urls),\n dates: createConfidenceField(null), // Simplified schema removed this for complexity reduction\n role: createConfidenceField(proj.role)\n })),\n certifications: (simplified.certifications || []).map((cert: any) => ({\n name: createConfidenceField(cert.name),\n issuer: createConfidenceField(cert.issuer),\n date: createConfidenceField(cert.date),\n expiryDate: createConfidenceField(cert.expiryDate),\n id: createConfidenceField(cert.id)\n })),\n languages: (simplified.languages || []).map((lang: any) => ({\n name: createConfidenceField(lang.name),\n proficiency: createConfidenceField(lang.proficiency)\n })),\n overallConfidence: simplified.overallConfidence || 0.7,\n missingFields: simplified.missingFields || [],\n detectedSections: simplified.detectedSections || []\n }\n\n return result\n}",[],[270,277],{"_103":285,"_95":286,"_155":287,"_157":271,"_158":140,"_159":288,"_161":117,"_105":289,"_100":196,"_163":177,"_165":290,"_166":22,"_167":-5,"_168":291,"_98":292},"ed28d142-2f6e-11f0-9767-569c3dd06744","resumeParser.ts","957203b2-2f17-11f0-8eca-569c3dd06744",57,["D",1747081392013],"import { \n ParsedResume, \n ConfidenceField, \n ResumeParserOptions,\n PersonalInfo,\n Skill,\n WorkExperience,\n Education,\n Project,\n Certification,\n Language,\n DateRange\n} from '../types/resume.ts';\nimport { generatePrompt } from '../prompts/resumePrompts.ts';\nimport { parseResumeWithAI, validateResumeWithAI, retryExtractionIfNeeded } from './openai.ts';\n\n/**\n * Helper function to create a confidence field with a value and confidence score\n */\nfunction createConfidenceField(\n value: T, \n confidence: number, \n standardization: string | null = null\n): ConfidenceField {\n return {\n value,\n confidence,\n standardization\n }\n}\n\n/**\n * Extract structured data from resume text with confidence scoring\n * \n * @param resumeText Plain text of the resume\n * @param options Optional parsing configuration\n * @returns Structured resume data with confidence scores\n */\nexport async function extractResumeData(\n resumeText: string,\n options?: ResumeParserOptions\n): Promise {\n try {\n console.log('Extracting data from resume text...');\n\n // Apply configuration options\n const confidenceThreshold = options?.confidenceThreshold || 0.3;\n const standardizationEnabled = options?.standardizationEnabled !== false; // Default to true\n\n // Step 1: Initial extraction with OpenAI\n console.log('Step 1: Performing initial extraction with OpenAI');\n let extractedData = await parseResumeWithAI(resumeText, options);\n\n if (!extractedData) {\n throw new Error('Failed to extract data from resume');\n }\n\n // Step 2: Retry extraction if initial results have issues\n console.log('Step 2: Checking extraction quality and retrying if needed');\n extractedData = await retryExtractionIfNeeded(extractedData, resumeText);\n\n // Step 3: Validate and standardize the data if enabled\n let processedData = extractedData;\n if (standardizationEnabled) {\n console.log('Step 3: Validating and standardizing extracted data');\n processedData = await validateResumeWithAI(extractedData);\n }\n\n // Step 4: Post-process the data (apply confidence thresholds, fill missing calculations)\n console.log('Step 4: Post-processing and final validation');\n const result = postProcessResumeData(processedData, confidenceThreshold);\n\n // Log completion status and confidence\n if (result.overallConfidence >= 0.7) {\n console.log('Resume extraction completed successfully with high confidence:', result.overallConfidence);\n } else if (result.overallConfidence >= 0.5) {\n console.log('Resume extraction completed with moderate confidence:', result.overallConfidence);\n } else {\n console.log('Resume extraction completed with low confidence:', result.overallConfidence);\n console.log('Missing fields:', result.missingFields);\n }\n\n return result;\n\n } catch (error) {\n console.error('Error extracting resume data:', error);\n throw new Error(`Failed to extract resume data: ${(error as Error).message}`);\n }\n}\n\n/**\n * Post-process resume data after AI extraction and validation\n *\n * @param data Extracted and validated resume data\n * @param confidenceThreshold Minimum confidence threshold for fields\n * @returns Finalized resume data\n */\nfunction postProcessResumeData(data: any, confidenceThreshold: number = 0.3): ParsedResume {\n try {\n // Ensure we have all required sections with at least empty arrays\n const processed: ParsedResume = {\n personalInfo: data.personalInfo || createDefaultPersonalInfo(),\n skills: data.skills || [],\n workExperience: data.workExperience || [],\n education: data.education || [],\n projects: data.projects || [],\n certifications: data.certifications || [],\n languages: data.languages || [],\n overallConfidence: data.overallConfidence || 0,\n missingFields: data.missingFields || [],\n detectedSections: data.detectedSections || []\n };\n\n // Calculate missing durations in work experience\n processed.workExperience = processed.workExperience.map(exp => {\n if (exp.dates) {\n // Calculate duration if not already provided\n if (!exp.dates.durationInMonths.value && exp.dates.startDate.value) {\n const startDate = new Date(exp.dates.startDate.value);\n const endDate = exp.dates.endDate.value ? new Date(exp.dates.endDate.value) : new Date();\n\n if (!isNaN(startDate.getTime()) && !isNaN(endDate.getTime())) {\n const months = calculateMonthsBetween(startDate, endDate);\n exp.dates.durationInMonths = createConfidenceField(months, 0.8);\n }\n }\n }\n return exp;\n });\n\n // Perform similar calculations for education dates\n processed.education = processed.education.map(edu => {\n if (edu.dates) {\n if (!edu.dates.durationInMonths.value && edu.dates.startDate.value) {\n const startDate = new Date(edu.dates.startDate.value);\n const endDate = edu.dates.endDate.value ? new Date(edu.dates.endDate.value) : new Date();\n\n if (!isNaN(startDate.getTime()) && !isNaN(endDate.getTime())) {\n const months = calculateMonthsBetween(startDate, endDate);\n edu.dates.durationInMonths = createConfidenceField(months, 0.8);\n }\n }\n }\n return edu;\n });\n\n // Calculate updated overall confidence score\n const confidenceValues: number[] = [];\n\n // Add personal info confidences\n Object.values(processed.personalInfo).forEach(field => {\n if (field && typeof field.confidence === 'number') {\n confidenceValues.push(field.confidence);\n }\n });\n\n // Add skills confidences\n processed.skills.forEach(skill => {\n confidenceValues.push(skill.name.confidence);\n });\n\n // Add work experience confidences (weighted higher)\n processed.workExperience.forEach(exp => {\n confidenceValues.push(exp.company.confidence * 1.5);\n confidenceValues.push(exp.title.confidence * 1.5);\n });\n\n // Add education confidences (weighted higher)\n processed.education.forEach(edu => {\n confidenceValues.push(edu.institution.confidence * 1.5);\n confidenceValues.push(edu.degree.confidence * 1.5);\n });\n\n // Calculate average confidence\n if (confidenceValues.length > 0) {\n processed.overallConfidence = confidenceValues.reduce((sum, val) => sum + val, 0) / confidenceValues.length;\n } else {\n processed.overallConfidence = 0;\n }\n\n // Round to 2 decimal places\n processed.overallConfidence = Math.round(processed.overallConfidence * 100) / 100;\n\n // Update missing fields by checking all required fields\n const updatedMissingFields: string[] = [];\n\n // Check personal info required fields\n const requiredPersonalFields = ['name', 'email', 'phone'];\n requiredPersonalFields.forEach(field => {\n const value = (processed.personalInfo as any)[field]?.value;\n if (!value || (processed.personalInfo as any)[field].confidence < confidenceThreshold) {\n updatedMissingFields.push(`personalInfo.${field}`);\n }\n });\n\n // Skills, work experience, education must have at least one entry\n if (processed.skills.length === 0) {\n updatedMissingFields.push('skills');\n }\n\n if (processed.workExperience.length === 0) {\n updatedMissingFields.push('workExperience');\n }\n\n if (processed.education.length === 0) {\n updatedMissingFields.push('education');\n }\n\n processed.missingFields = updatedMissingFields;\n\n return processed;\n } catch (error) {\n console.error('Error in post-processing resume data:', error);\n // Return original data if post-processing fails\n return data as ParsedResume;\n }\n}\n\n/**\n * Create a default personal info object with empty fields\n */\nfunction createDefaultPersonalInfo(): PersonalInfo {\n return {\n name: createConfidenceField('', 0),\n email: createConfidenceField('', 0),\n phone: createConfidenceField('', 0),\n location: createConfidenceField(null, 0),\n linkedin: createConfidenceField(null, 0),\n github: createConfidenceField(null, 0),\n website: createConfidenceField(null, 0),\n summary: createConfidenceField(null, 0)\n };\n}\n\n/**\n * Calculate months between two dates\n */\nfunction calculateMonthsBetween(startDate: Date, endDate: Date): number {\n return (endDate.getFullYear() - startDate.getFullYear()) * 12 +\n (endDate.getMonth() - startDate.getMonth());\n}\n\n\n/**\n * Calculate confidence score for a field based on various factors\n * \n * @param value The field value\n * @param fieldType The type of field being scored\n * @param context Additional context for scoring\n * @returns Confidence score between 0 and 1\n */\nexport function calculateConfidence(\n value: any, \n fieldType: string, \n context?: Record\n): number {\n // In the real implementation, this would calculate confidence based on:\n // - Field completion/presence\n // - Format validity\n // - Consistency with other fields\n // - Pattern matching against known formats\n // - Etc.\n \n // For now, return a placeholder score\n if (!value) return 0;\n return 0.8;\n}\n\n/**\n * Detect and standardize common terminology\n * \n * @param value The original value\n * @param fieldType The type of field to standardize\n * @returns Standardized value and standardization notes\n */\nexport function standardizeTerminology(\n value: string, \n fieldType: 'degree' | 'jobTitle' | 'skill' | 'language' | string\n): { standardized: string; notes: string | null } {\n // This would contain logic to standardize different types of fields\n // For example, converting \"BS\" to \"Bachelor of Science\"\n \n // For now, return the original value\n return { standardized: value, notes: null };\n}\n\n/**\n * Add confidence scores to all fields in the resume data\n * \n * @param data Raw resume data without confidence scores\n * @returns Resume data with confidence scores added\n */\nexport function addConfidenceScores(data: any): ParsedResume {\n // This function would iterate through the resume data and add\n // confidence scores to each field based on various factors\n \n // For now, return a placeholder response\n return data as ParsedResume;\n}",[],[270,286],[270],{"_103":295,"_95":296,"_155":297,"_157":-5,"_158":140,"_159":298,"_161":117,"_105":299,"_100":32,"_163":300,"_165":301,"_166":22,"_167":-5,"_168":302,"_98":303},"9543d096-2f17-11f0-8234-569c3dd06744",".vtignore","9543d168-2f17-11f0-8234-569c3dd06744",6,["D",1747043878343],"file",".git\n.vscode\n.cursorrules\n.DS_Store\nnode_modules\nvendor",[],[296],{"_103":305,"_95":306,"_155":307,"_157":-5,"_158":140,"_159":308,"_161":117,"_105":309,"_100":32,"_163":300,"_165":310,"_166":22,"_167":-5,"_168":311,"_98":312},"95777626-2f17-11f0-8670-569c3dd06744","deno.json","95777824-2f17-11f0-8670-569c3dd06744",12,["D",1747043878671],"{\n \"$schema\": \"https://raw.githubusercontent.com/denoland/deno/348900b8b79f4a434cab4c74b3bc8d4d2fa8ee74/cli/schemas/config-file.v1.json\",\n \"lock\": false,\n \"compilerOptions\": {\n \"noImplicitAny\": false,\n \"strict\": false,\n \"types\": [\n \"https://www.val.town/types/valtown.d.ts\"\n ],\n \"lib\": [\n \"dom\",\n \"dom.iterable\",\n \"dom.asynciterable\",\n \"deno.ns\",\n \"deno.unstable\"\n ]\n },\n \"lint\": {\n \"files\": {\n \"include\": [\n \"deno:/https/esm.town/**/*\"\n ]\n },\n \"rules\": {\n \"exclude\": [\n \"no-explicit-any\"\n ]\n }\n },\n \"node_modules_dir\": false,\n \"experimental\": {\n \"unstable-node-globals\": true,\n \"unstable-temporal\": true,\n \"unstable-worker-options\": true,\n \"unstable-sloppy-imports\": true\n }\n}",[],[306],{"_103":314,"_95":315,"_155":115,"_157":-5,"_158":140,"_159":316,"_161":117,"_105":317,"_100":318,"_163":319,"_165":320,"_166":22,"_167":321,"_168":322,"_98":323},"b901fe30-4854-11f0-aae2-76b3cceeab13","index.ts",61,["D",1749818916673],10,"http","import {OpenAPIHono} from 'npm:@hono/zod-openapi'\nimport {cors} from 'npm:hono/cors'\nimport {logger} from 'npm:hono/logger'\nimport {swaggerUI} from 'npm:@hono/swagger-ui'\nimport {errorHandler} from './middleware/errorHandler.ts'\nimport resumeRouter from './controllers/resumeController.ts'\nimport healthRouter from './controllers/healthController.ts'\nimport mcpController from \"./controllers/mcpController.ts\"\n\n// Create OpenAPIHono app\nconst app = new OpenAPIHono()\n\n// Global middleware\napp.use('*', logger())\napp.use('*', cors({\n origin: '*',\n allowMethods: ['GET', 'POST', 'OPTIONS'],\n allowHeaders: ['Content-Type', 'Authorization'],\n exposeHeaders: ['Content-Length'],\n maxAge: 600,\n credentials: true,\n}))\n\n// Unwrap Hono errors to see original error details\napp.onError((err, c) => {\n return errorHandler(err, c)\n})\n\n// Mount routers\napp.route('/', healthRouter)\napp.route('/api', resumeRouter)\napp.route('/mcp', mcpController)\n// OpenAPI documentation\napp.doc('/doc', {\n openapi: '3.0.0',\n info: {\n title: 'Resume Parser API',\n version: '1.0.0',\n description: 'API for parsing resumes and extracting structured data with confidence scoring'\n },\n servers: [\n {\n url: 'https://prashamhtrivedi-resumeparser.val.run',\n description: 'Resume Praser APIs'\n }\n ]\n})\n\n// Use the middleware to serve Swagger UI at /api-doc\napp.get('/api-doc', swaggerUI({url: '/doc'}))\n\n// Not found handler\napp.notFound((c) => {\n return c.json({\n success: false,\n error: {\n message: 'Resource not found',\n code: 'NOT_FOUND',\n }\n }, 404)\n})\n\n// This is the entry point for HTTP vals\nexport default app.fetch","prashamhtrivedi-resumeparser",[],[315],"files",[152,170,181,190,201,207,217,223,233,240,251,258,268,275,284,294,304,313],"counts",{"_328":32,"_329":176,"_330":32,"_331":332},"pulls","branches","forks","envVars",{"_333":176,"_334":32},"count","emptyCount","contents",[],{"_103":224,"_95":225,"_155":226,"_157":220,"_158":140,"_159":227,"_161":117,"_105":338,"_100":160,"_163":177,"_165":229,"_166":22,"_167":-5,"_98":339},["D",1747076851535],[219,225],"upstreamProject","upstreamBranch","emptyEnvVars","isNew","flags",[],"routes/_app.x.$handle.$name.branch.$branch.history_.$",{"_93":94,"_95":96,"_97":43,"_98":99,"_101":348,"_300":353,"_356":357,"_369":370,"_375":-7,"_376":32,"_377":22},{"_103":104,"_95":96,"_105":349,"_107":-5,"_108":22,"_109":110,"_111":72,"_112":-5,"_14":113,"_114":115,"_116":117,"_118":119,"_120":32,"_16":121,"_122":123,"_124":123,"_125":126,"_127":126,"_128":126,"_129":123,"_130":123,"_23":350,"_134":351,"_137":352,"_149":22},["D",1747042122266],{"_93":94,"_132":133},{"_93":94,"_136":133},{"_100":139,"_95":43,"_103":140,"_141":104,"_142":-5,"_143":-5,"_144":-5,"_145":-5,"_146":-5,"_147":96,"_148":94},{"_103":224,"_95":225,"_155":226,"_157":220,"_158":140,"_159":227,"_161":117,"_105":354,"_100":160,"_163":177,"_165":229,"_166":22,"_167":-5,"_98":355},["D",1747076851535],[219,225],"history",[358,360,365],{"_103":224,"_158":140,"_155":226,"_95":225,"_165":229,"_163":177,"_100":160,"_157":220,"_105":359,"_159":227,"_167":-5},["D",1747076851535],{"_103":361,"_158":140,"_155":226,"_95":225,"_165":362,"_163":177,"_100":176,"_157":220,"_105":363,"_159":364,"_167":-5},"44fad4b4-2f25-11f0-aad8-569c3dd06744","/**\n * Generate prompts for different resume parsing operations\n * \n * @param type The type of prompt to generate\n * @returns The prompt text\n */\nexport function generatePrompt(type: 'extract' | 'validate' | 'confidence'): string {\n switch (type) {\n case 'extract':\n return `Extract structured data from the following resume with HIGH PRECISION.\n\nYou MUST return a JSON object with the following structure:\n{\n \"personalInfo\": {\n \"name\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"email\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"phone\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"location\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"linkedin\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"github\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"website\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"summary\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null }\n },\n \"skills\": [\n {\n \"name\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"category\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"proficiency\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"yearsOfExperience\": { \"value\": number | null, \"confidence\": number, \"standardization\": string | null }\n }\n ],\n \"workExperience\": [\n {\n \"company\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"title\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"location\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"dates\": {\n \"startDate\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"endDate\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"durationInMonths\": { \"value\": number | null, \"confidence\": number, \"standardization\": string | null },\n \"current\": { \"value\": boolean, \"confidence\": number, \"standardization\": string | null }\n },\n \"responsibilities\": { \"value\": string[], \"confidence\": number, \"standardization\": string | null },\n \"technologies\": { \"value\": string[], \"confidence\": number, \"standardization\": string | null },\n \"achievements\": { \"value\": string[], \"confidence\": number, \"standardization\": string | null }\n }\n ],\n \"education\": [\n {\n \"institution\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"degree\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"field\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"dates\": {\n \"startDate\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"endDate\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"durationInMonths\": { \"value\": number | null, \"confidence\": number, \"standardization\": string | null },\n \"current\": { \"value\": boolean, \"confidence\": number, \"standardization\": string | null }\n },\n \"gpa\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"coursework\": { \"value\": string[] | null, \"confidence\": number, \"standardization\": string | null },\n \"achievements\": { \"value\": string[] | null, \"confidence\": number, \"standardization\": string | null }\n }\n ],\n \"projects\": [\n {\n \"name\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"description\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"technologies\": { \"value\": string[], \"confidence\": number, \"standardization\": string | null },\n \"urls\": { \"value\": string[] | null, \"confidence\": number, \"standardization\": string | null },\n \"dates\": { \n \"value\": {\n \"startDate\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"endDate\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"durationInMonths\": { \"value\": number | null, \"confidence\": number, \"standardization\": string | null },\n \"current\": { \"value\": boolean, \"confidence\": number, \"standardization\": string | null }\n } | null, \n \"confidence\": number, \n \"standardization\": string | null \n },\n \"role\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null }\n }\n ],\n \"certifications\": [\n {\n \"name\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"issuer\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"date\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"expiryDate\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null },\n \"id\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null }\n }\n ],\n \"languages\": [\n {\n \"name\": { \"value\": string, \"confidence\": number, \"standardization\": string | null },\n \"proficiency\": { \"value\": string | null, \"confidence\": number, \"standardization\": string | null }\n }\n ],\n \"overallConfidence\": number,\n \"missingFields\": string[],\n \"detectedSections\": string[]\n}\n\nIMPORTANT GUIDELINES:\n1. Confidence scores range from 0 to 1, where:\n - 0.9-1.0: Very high confidence (clear, unambiguous data)\n - 0.7-0.9: High confidence (likely correct, but could have minor issues)\n - 0.5-0.7: Medium confidence (possibly correct, but some uncertainty)\n - 0.3-0.5: Low confidence (significant uncertainty)\n - 0.0-0.3: Very low confidence (likely incorrect or missing data)\n\n2. Standardization notes should be provided whenever you transform or normalize the data, such as:\n - Converting abbreviations to full forms (e.g., \"BS\" → \"Bachelor of Science\")\n - Normalizing job titles (e.g., \"Dev\" → \"Developer\")\n - Standardizing technology names (e.g., \"JS\" → \"JavaScript\")\n - Date format standardization\n - If no standardization was performed, use null\n\n3. For missing or unavailable data:\n - Use null for string fields that are missing\n - Use empty array [] for array fields that are missing\n - Use 0 as the confidence score for missing fields\n - Add the field path to missingFields array (e.g., \"personalInfo.website\")\n\n4. For dates:\n - Use ISO format (YYYY-MM) whenever possible\n - Mark current positions with current: true and null endDate\n - Calculate durationInMonths when possible\n\n5. For skills:\n - Infer categories when possible (e.g., \"Programming Languages\", \"Tools\", \"Frameworks\")\n - Infer proficiency when mentioned (\"Beginner\", \"Intermediate\", \"Advanced\", \"Expert\")\n - Detect years of experience when mentioned\n\n6. For work experience:\n - Extract technologies used in each role when mentioned\n - Separate responsibilities from achievements when possible\n - Standardize job titles to industry norms\n\n7. DO NOT fabricate or assume information not present in the resume. If information is missing, use null with a confidence score of 0.\n\n8. The overallConfidence score should be a weighted average of all fields, giving more weight to critical fields like name, contact info, work experience, and education.\n\n9. The detectedSections array should list all major sections found in the resume (e.g., personal_information, skills, experience, education, etc.).`;\n \n case 'validate':\n return `Validate the following resume data for consistency, accuracy, and standardization.\n\nFor each field in the data:\n1. Check for inconsistencies and validate format\n2. Standardize terminology where appropriate:\n - Degree names (e.g., \"BS\" → \"Bachelor of Science\", \"MS\" → \"Master of Science\")\n - Job titles (e.g., \"SWE\" → \"Software Engineer\", \"PM\" → \"Product Manager\")\n - Technology names (e.g., \"JS\" → \"JavaScript\", \"TS\" → \"TypeScript\")\n - Institution names (e.g., \"NYU\" → \"New York University\")\n3. Validate dates:\n - Ensure date ranges are logical (start before end)\n - Verify that duration calculations are accurate\n - Check for overlapping or inconsistent date ranges\n4. Look for contradictions:\n - Skills mentioned in work experience but not in skills section\n - Technologies that don't match the responsibilities\n - Date inconsistencies between sections\n5. Update confidence scores:\n - Increase confidence for validated fields\n - Decrease confidence for fields with detected issues\n - Add standardization notes when changes are made\n\nReturn the validated and standardized data in the exact same format, including:\n- All confidence scores updated based on validation\n- Standardization notes added for any changes made\n- Overall confidence score recalculated\n- Missing fields list updated if additional missing fields are found\n\nIf you detect serious inconsistencies or potential fabrications, mark those fields with lower confidence scores and add appropriate standardization notes.`;\n \n case 'confidence':\n return `Analyze the following resume data and calculate confidence scores for each field.\n\nConfidence scoring criteria by field type:\n\n1. Personal Information:\n - Name: Check for full name vs. partial, common formats\n - Email: Validate format, check for standard patterns\n - Phone: Validate format, check country code presence\n - Location: Check for specificity (city/state/country)\n - Social/Web: Validate URL formats and completeness\n\n2. Skills:\n - Compare against known technology/skill names\n - Check for specificity (general vs. specific skills)\n - Look for proficiency indicators\n - Consider consistency with work experience\n\n3. Work Experience:\n - Validate company names against known entities\n - Check job title standardization\n - Verify date formats and calculations\n - Evaluate responsibility descriptions (specific vs. vague)\n - Look for quantifiable achievements\n - Check for technology mentions that match skills\n\n4. Education:\n - Validate institution names\n - Check degree nomenclature standardization\n - Verify field of study specificity\n - Validate date formats and ranges\n - Check GPA format and range\n - Evaluate coursework relevance\n\n5. Projects & Certifications:\n - Evaluate description specificity\n - Check URL validity\n - Verify technology consistency with skills\n - Validate certification issuers\n - Check date formats and validity\n\nRate each field on a scale from 0 to 1, where:\n- 0.9-1.0: Very high confidence (clear, unambiguous data)\n- 0.7-0.9: High confidence (likely correct, but could have minor issues)\n- 0.5-0.7: Medium confidence (possibly correct, but some uncertainty)\n- 0.3-0.5: Low confidence (significant uncertainty)\n- 0.0-0.3: Very low confidence (likely incorrect or missing data)\n\nFor any field where confidence is below 0.7, include specific reasons in the standardization notes.\n\nReturn the data with confidence scores added to each field and an overall confidence score that is a weighted average of all fields.`;\n \n default:\n return '';\n }\n}",["D",1747049756595],21,{"_103":366,"_158":140,"_155":226,"_95":225,"_165":367,"_163":177,"_100":32,"_157":220,"_105":368,"_159":246,"_167":-5},"954eb6c8-2f17-11f0-8670-569c3dd06744","/**\n * Generate prompts for different resume parsing operations\n * \n * @param type The type of prompt to generate\n * @returns The prompt text\n */\nexport function generatePrompt(type: 'extract' | 'validate' | 'confidence'): string {\n switch (type) {\n case 'extract':\n return `Extract structured data from the following resume. \nReturn a JSON object with the following sections:\n- contactDetails: name, email, phone, linkedin, github, website\n- skills: array of skill categories each with a name and items array\n- workExperience: array of positions with company, role, duration, responsibilities\n- education: array of education entries with degree, institution, location, duration\n\nFor each field, include a confidence score from 0-1 indicating your certainty.\nStandardize degree names to common formats (e.g., \"BS\" to \"Bachelor of Science\").\nDO NOT fabricate or assume data not present in the resume.`;\n \n case 'validate':\n return `Validate the following resume data for consistency and accuracy.\nIdentify any fields that seem incorrect or inconsistent and flag them.\nStandardize terminology where appropriate (e.g., degree names, job titles).\nReturn the validated and standardized data in the same format.`;\n \n case 'confidence':\n return `Analyze the following resume data and add confidence scores for each field.\nRate each field on a scale from 0 to 1, where:\n- 0.9-1.0: Very high confidence (clear, unambiguous data)\n- 0.7-0.9: High confidence (likely correct, but could have minor issues)\n- 0.5-0.7: Medium confidence (possibly correct, but some uncertainty)\n- 0.3-0.5: Low confidence (significant uncertainty)\n- 0.0-0.3: Very low confidence (likely incorrect or missing data)\nReturn the data with confidence scores added to each field.`;\n \n default:\n return '';\n }\n}",["D",1747043878346],"versions",[371,373,374],{"_300":358,"_372":360},"previous",{"_300":360,"_372":365},{"_300":365,"_372":-5},"versionNumber","page","hasNext","actionData","errors"]

prashamtrivedi

resume-parser

History for resumePrompts.ts