Lexi - AI Legal Assistant

[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10},"actionData","errors","root",{"_162":163,"_164":165},"routes/_app.x.$handle.$name",{"_11":12,"_13":14,"_15":-7,"_61":132,"_16":-7,"_133":-5,"_17":134,"_135":136,"_28":137,"_21":138,"_23":139,"_140":-5,"_141":-5,"_142":-5,"_143":35,"_144":145,"_146":38,"_147":-5,"_148":-5,"_149":150},"routes/_app.x.$handle.$name._index",{"_11":12,"_13":14,"_15":-7,"_16":-7,"_17":18,"_19":20,"_21":22,"_23":24,"_25":-7,"_26":-5,"_27":-5,"_28":29,"_30":31,"_32":-5,"_33":-5,"_34":35,"_36":35,"_37":38},"handle","legal","name","lexi","branchName","version","project",{"_39":78,"_13":14,"_50":79,"_80":-5,"_81":35,"_82":83,"_84":83,"_85":-5,"_86":-5,"_87":88,"_89":43,"_90":38,"_91":38,"_92":-5,"_93":67,"_94":95,"_96":-5,"_97":-5,"_98":-5,"_99":-5,"_100":101,"_102":101,"_103":104,"_105":104,"_106":104,"_107":101,"_108":101,"_109":110,"_111":112,"_113":114,"_115":35},"tags",[],"counts",{"_66":67,"_68":69,"_70":67,"_71":72,"_73":74},"contents",[63],"readme","upstream","pull","files",[31],"preview",{"_39":40,"_13":41,"_42":43,"_44":-5,"_45":46,"_47":48,"_49":38,"_50":51,"_16":52,"_53":54,"_55":56,"_57":35,"_58":-5,"_59":60,"_61":62},"ahead","behind","showHomeCustomizationHint",false,"showHTTPPreviewSettingsHint","hasHTTPFiles",true,"id","a489ae08-4494-11f0-8173-76b3cceeab13","main.tsx","fileId","c7323414-41cb-11f0-a13f-76b3cceeab13","parentId","branchId","c7309348-41cb-11f0-a13f-76b3cceeab13","branchVersion",25,"latest","createdAt",["D",1749406565467],24,"fileType","http","content","const legalTaskSuggestionSystemPromptTemplate = `\nAI legal document review assistant. Your goal is to suggest a comprehensive and diverse list of potential legal tasks or analyses that a user might want toperform on the provided document text. Return your suggestions STRICTLY as a JSON object. The JSON object's keys are category names (e.g., \"Information Extraction\", \"Risk Assessment\"). Values are arrays of objects, each with \"task\" (a string describing the task) and \"is_high_priority\" (boolean). Aim for 3-5 categories, each with 3-7 relevant tasks. Mark foundational tasks as \"is_high_priority: true\".\n\nDOCUMENT TEXT:\n==============\n%%DOCUMENT_TEXT%%\n`;\n\nconst legalAIAnalysisSystemPromptTemplate = `\nAI System for Precise Legal Information Extraction. Your goal is to perform the user-specified legal task(s) on the provided legal text. Your output MUST be a single JSON object (no markdown formatting) that strictly follows the schema provided below.\nLEGAL TASK(S) / QUERY: %%LEGAL_TASK_QUERY%%\nDOCUMENT SOURCE: %%INPUT_SOURCE_DESCRIPTION%%\n\nAnalyze the document text provided in the user message. Meticulously map ALL findings to the JSON schema, presenting information objectively. If the text is insufficient or ambiguous for parts of the query, note this in the 'limitations_or_warnings' field.\n\nOutput JSON Schema:\n{\n \"request_details\": { \"operation_performed\": \"string\", \"input_source_description\": \"string\" },\n \"analysis_summary\": { \"overview\": \"string\", \"key_findings_count\": \"integer\" },\n \"detailed_findings\": [\n {\n \"finding_id\": \"string\", // e.g., \"FIN-001\"\n \"finding_type\": \"string\", // e.g., \"Date Extraction\"\n \"description\": \"string\",\n \"severity_or_importance\": \"string\", // e.g., \"High\", \"Medium\", \"Low\"\n \"relevant_input_excerpt\": \"string\",\n \"page_or_section_reference\": \"string\",\n \"implications\": \"string\",\n \"confidence_score\": \"float\" // 0.0 to 1.0\n }\n ],\n \"execution_metadata\": {\n \"status\": \"string\", // e.g., \"Completed\", \"Completed with warnings\"\n \"timestamp\": \"string\", // ISO 8601\n \"limitations_or_warnings\": \"string\",\n \"disclaimer\": \"string\"\n }\n}\n`;\n\n// --- Type Definitions ---\ninterface LogEntry {\n agent: string;\n message: string | object;\n type: \"step\" | \"result\" | \"error\" | \"input\" | \"final\" | \"info\";\n}\ninterface LegalAIResponse {\n request_details: { operation_performed: string; input_source_description: string };\n analysis_summary: { overview: string; key_findings_count: number };\n detailed_findings: Array<{\n finding_id: string;\n finding_type: string;\n description: string;\n severity_or_importance?: string;\n relevant_input_excerpt?: string;\n page_or_section_reference?: string;\n implications?: string;\n confidence_score?: number;\n }>;\n execution_metadata: { status: string; timestamp: string; limitations_or_warnings: string; disclaimer: string };\n}\ninterface LexiDocument {\n id: string;\n name: string;\n text: string;\n sourceType: \"file\";\n sourceValue: string;\n dateAdded: string;\n type: \"document\";\n}\ninterface LexiAnalysis {\n id: string;\n docId: string;\n docName?: string;\n query: string;\n result: LegalAIResponse;\n datePerformed: string;\n type: \"analysis\";\n}\ninterface GroupedSuggestions {\n [category: string]: { task: string; is_high_priority: boolean }[];\n}\n\nfunction generateHtmlShell(initialQuery, sourceUrl) {\n const escapedQuery = initialQuery.replace(//g, \">\");\n return `\n\n\n\n\n\nLexi - AI Legal Assistant\n\n\n\n

\n\n

\n${\n [\"dashboard\", \"new-lex\", \"history\", \"settings\", \"document-detail\", \"analysis-detail\"].map(v =>\n `

`\n ).join(\"\")\n }\n

\n\n\n`;\n}\n\nexport default async function(req: Request) {\n const { OpenAI } = await import(\"https://esm.town/v/std/openai\");\n const { PDFExtract } = await import(\"npm:pdf.js-extract\");\n\n const CORS_HEADERS = {\n \"Access-Control-Allow-Origin\": \"*\",\n \"Access-Control-Allow-Methods\": \"POST, GET, OPTIONS\",\n \"Access-Control-Allow-Headers\": \"Content-Type, Accept\",\n };\n if (req.method === \"OPTIONS\") return new Response(null, { status: 204, headers: CORS_HEADERS });\n\n const url = new URL(req.url);\n const format = url.searchParams.get(\"format\");\n const action = url.searchParams.get(\"action\");\n const sourceUrl = import.meta.url.replace(\"esm.town\", \"val.town\");\n const openai = new OpenAI();\n const MAX_TEXT_SUGGEST = 20000;\n const MAX_TEXT_ANALYZE = 30000;\n\n async function extractPdfText(data: ArrayBuffer, fileName: string, log: LogEntry[]): Promise {\n log.push({ agent: \"PDF\", type: \"step\", message: `Processing: ${fileName}` });\n try {\n const pdfExtract = new PDFExtract();\n const pdfData = await pdfExtract.extractBuffer(data, {});\n const text = pdfData.pages.map(page => page.content.map(item => item.str).join(\" \")).join(\"\\n\\n\");\n if (!text?.trim()) {\n log.push({ agent: \"PDF\", type: \"error\", message: \"Extraction yielded empty text.\" });\n return null;\n }\n log.push({ agent: \"PDF\", type: \"info\", message: `Extracted ~${text.length} chars.` });\n return text;\n } catch (error) {\n log.push({ agent: \"PDF\", type: \"error\", message: `Extraction error: ${error.message}` });\n console.error(\"PDF Extraction Error:\", error);\n return null;\n }\n }\n\n async function callAI(\n systemPrompt: string,\n userMessage: string,\n log: LogEntry[],\n agentName: string,\n ): Promise