Janady13/missionmatch / frontend/src/lib/resume.ts

Source identity

Field	Recorded value
Source type	GitHub source
Original record	Janady13/missionmatch / frontend/src/lib/resume.ts
Date	Not specified in source
Repository	Janady13/missionmatch
Revision	720d47fae476b0d38477ae4401697d372b1ea2bc
Branch	feature/opportunity-to-outcome
Path or archive member	frontend/src/lib/resume.ts
Original SHA-256	2f6d6549cf987a90eb05e50e2146264d4e070b1d8b8e0f252ef8537b1ef4dae2
Git blob SHA	8c6853c71677ac2e0aadc9311c72d022f74c97d8
Reading-copy SHA-256	2f6d6549cf987a90eb05e50e2146264d4e070b1d8b8e0f252ef8537b1ef4dae2

Original source ↗

Every recorded occurrence

These 35 records contain identical original source bytes. Their paths and revisions remain part of the archive.

Original record	Branch or archive	Revision / identifier	Original location
Janady13/missionmatch / frontend/src/lib/resume.ts	feature/opportunity-to-outcome	720d47fae476b0d38477ae4401697d372b1ea2bc	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	per-tenant-routing	9bde0f9a74b89d01980fa658e5253acda7528d0c	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	collision-deconfliction	4ee5a09972608ed89f8377c6665542d5f450f0a6	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	ws6-retrieval	6c56e29231ec05735927fa265c97d517499dfeaf	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	workstream-a-shared-core	e0de03ac3a1ec826bd15d0e048e3b02baa23eeff	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	m5-price-to-win	41000ad30fff4129fdaf5ec84e15ba14c27d4890	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	workstream-c-slatematch	5cd60c918301b46ecd6ed3e03e20c395da0062e2	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	ws1-ontology-sync	b596f449a32a1016fbd403c36505cc00a1ac9ee6	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	master	1e9249840a4302ec82e123b9c9cb6f75badea3e4	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	ws8-teaching-loop	c823f0fb6bb2aba0f39001eb6de96f4000d9321f	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	w3-ai-staffing-recommendations	43d56387270ef4da4cabd794d97ca6a49909bc7d	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	mag-substrate	b82c7100a492dec67dfd1eda6b3f53def859be35	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	ws3-clearance	4b8e7ce594c068621867f9c9deed0ec64ae6c553	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	workstream-d-command-center	255e7d1d420abf5e277d0cd8c9b14f6f72b6acb9	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	phase2-eval-fairness	d361af254701975fe239927656ababb56374655e	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	workstream-e-dataflow-deploy	433815a042f9741411e60fa553fe72143e9b10db	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	ruthless-debug-sweep	b7d742f8f032984be3817efffa3eb8b13b651c1f	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	time-to-fill-survival	b66e1e505863d4f17741a4b9ede73509e45bd6c6	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	ai-intelligence-layer	5d5ddff70393a197c5f54a45280bb5dd42b24755	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	graphrag-foundation	b0ddeb9f633fd138d09309d1a17b7d13bc3a67f7	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	workstream-a-opportunity-intelligence	1675031db74316e100782d063e9eda1a597fe3a1	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	deploy-and-ops	bdacd9d0b4d5998d25fc74bf4007f4c3ec538c7a	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	ws9-bench-ingestion	f585a9ef32fa554663ce1fba9a414a54b6e49cf0	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	ws2-alias-relationship	e4dada4601c6eb101830f9d223a9655dfa120616	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	presentation-readiness	eb574acaa29d71ce4d0f852572e87c275c79fdb7	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	vendor-intelligence	fe9796a19c261951d3f2565206c0c1b5ac294465	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	workstream-b-screenmatch	09015496d565a172bcaa15cfe9a8e4b072b13df0	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	ws7-federal-signal	6a9f6ce9cfb13345ffb47770d4960e7c2d9c17c2	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	workstream-e-commercial-stubs	4437da660539284a400ee380b9b427a300c369ee	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	ai-brain-live	aac32fc49c878cc5bca099459307611562d0bac1	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	workstream-d-missionmatch	3be46ba89edb737182075a2b320509b42dc916f8	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	phase2-connectors	e038d4c687aec16f8bd9bc4d907eea14d91f8704	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	workstream-b-sam-contract-awards	d5101cceb98e70000cebb31d3e7ac73d747d3a6d	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	feature/aioe-upgrade	2178e7298d0e5d9af83afa12402fd587f0ba6ed8	Open original
Janady13/missionmatch / frontend/src/lib/resume.ts	capture-readiness	bae6af7cb2997e819174a9e9cfff5e16067a1975	Open original

Recovered source transcript

Download complete source text ↓121 lines · 4.5 KB

// resume.ts extracts plain TEXT from a PDF or DOCX file ENTIRELY IN THE BROWSER. The file bytes never
// leave the tab and are never written to storage; only the extracted text is later POSTed to our own
// authed API for structured extraction and scoring. This keeps resume bytes (potential CUI) off the
// wire and needs no API key, so nothing here can leak a credential into the bundle (CLAUDE.md rule 3).
import * as pdfjsLib from 'pdfjs-dist'
import workerUrl from 'pdfjs-dist/build/pdf.worker.min.mjs?url'
import mammoth from 'mammoth'

// Resolve the pdf.js worker from the SAME installed package via Vite's ?url, so the worker version
// always matches the API version (a version mismatch is the most common pdf.js failure) and the worker
// is self hosted as its own chunk, never a CDN.
pdfjsLib.GlobalWorkerOptions.workerSrc = workerUrl

export type ExtractedSource = 'pdf' | 'docx'

export interface Extracted {
  text: string
  source: ExtractedSource
  // lowConfidence is set when almost no selectable text was found, the signature of a scanned or image
  // only document. The caller then routes the user to the paste text fallback rather than scoring noise.
  lowConfidence: boolean
}

// ResumeParseError carries a user facing message; the modal shows it and offers the paste fallback.
export class ResumeParseError extends Error {}

const PDF_MAX = 15 * 1024 * 1024
const DOCX_MAX = 10 * 1024 * 1024
const LOW_CONFIDENCE_CHARS = 80

// sniff classifies the file by its magic bytes, not its extension: PDF starts with %PDF, a DOCX is a
// zip and starts with PK\x03\x04. A renamed file is rejected rather than mis parsed.
function sniff(bytes: Uint8Array): ExtractedSource | null {
  if (bytes.length >= 4 && bytes[0] === 0x25 && bytes[1] === 0x50 && bytes[2] === 0x44 && bytes[3] === 0x46) {
    return 'pdf'
  }
  if (bytes.length >= 4 && bytes[0] === 0x50 && bytes[1] === 0x4b && bytes[2] === 0x03 && bytes[3] === 0x04) {
    return 'docx'
  }
  return null
}

// extractResumeText reads the file in memory, validates it, and returns the extracted text. It throws a
// ResumeParseError with a friendly message on anything it cannot handle.
export async function extractResumeText(file: File): Promise<Extracted> {
  if (file.size === 0) {
    throw new ResumeParseError('That file is empty.')
  }
  const buf = await file.arrayBuffer()
  const bytes = new Uint8Array(buf)
  const kind = sniff(bytes)
  if (kind === null) {
    throw new ResumeParseError('Unsupported file. Upload a PDF or DOCX resume, or paste the text.')
  }
  if (kind === 'pdf') {
    if (file.size > PDF_MAX) {
      throw new ResumeParseError('That PDF is too large. The limit is 15 MB.')
    }
    return extractPdf(bytes)
  }
  if (file.size > DOCX_MAX) {
    throw new ResumeParseError('That DOCX is too large. The limit is 10 MB.')
  }
  return extractDocx(buf)
}

async function extractPdf(bytes: Uint8Array): Promise<Extracted> {
  const loadingTask = pdfjsLib.getDocument({ data: bytes })
  let doc
  try {
    doc = await loadingTask.promise
  } catch {
    throw new ResumeParseError('Could not read that PDF. It may be encrypted or corrupt. Paste the text instead.')
  }
  try {
    const parts: string[] = []
    for (let i = 1; i <= doc.numPages; i++) {
      const page = await doc.getPage(i)
      const content = await page.getTextContent()
      for (const item of content.items) {
        if ('str' in item) {
          parts.push(item.str)
          if (item.hasEOL) parts.push('\n')
        }
      }
      parts.push('\n')
      page.cleanup()
    }
    const text = normalize(parts.join(' '))
    return { text, source: 'pdf', lowConfidence: nonWhitespace(text) < LOW_CONFIDENCE_CHARS }
  } finally {
    // Releases the document and tears down the worker resources for this parse.
    await loadingTask.destroy()
  }
}

async function extractDocx(buf: ArrayBuffer): Promise<Extracted> {
  let value: string
  try {
    const result = await mammoth.extractRawText({ arrayBuffer: buf })
    value = result.value
  } catch {
    throw new ResumeParseError('Could not read that DOCX. Paste the text instead.')
  }
  const text = normalize(value)
  return { text, source: 'docx', lowConfidence: nonWhitespace(text) < LOW_CONFIDENCE_CHARS }
}

// normalize collapses runs of spaces and tabs but keeps newlines, which the name detector relies on.
function normalize(s: string): string {
  return s.replace(/[ \t]+/g, ' ').replace(/\n{3,}/g, '\n\n').trim()
}

function nonWhitespace(s: string): number {
  let n = 0
  for (const c of s) {
    if (!/\s/.test(c)) n++
  }
  return n
}