Skip to content

docs: give the README a lead, badges, and a quick start it was missing - #19

Merged
bibeshpyakurel merged 1 commit into
mainfrom
docs/readme-refresh
Sep 21, 2026
Merged

bibeshpyakurel merged 1 commit into
mainfrom
docs/readme-refresh

Conversation

@bibeshpyakurel

Copy link
Copy Markdown
Owner

The README opened with three paragraphs of prose and never said how to run the project. This adds the missing front matter and the missing Quick start.

Added: a one-line description, CodeQL / license / top-language badges (each verified 200), an explicit note that there is no hosted demo because the dataset is imported into local SQLite, a Screenshots placeholder section, What it demonstrates technically, a Quick start with an API endpoint table, Status and License. The existing prose is kept and reordered so Data Source precedes Quick start.

Note on sourcing: VisaTrack-Pro has no entry in the master documents repo — it is absent from the career document §5 and from all five industry resumes. Every claim here is therefore taken from the repository itself:

  • USCIS schema normalization across fiscal years — backend/scripts/importCSV.js, syncUSCIS.js
  • Source-level row storage and the processed-vs-stored metric split — the existing README's own Data Pipeline Design section
  • node-cron scheduling and sync-run metadata — backend/services/dataSync.js, USCIS_SYNC_SCHEDULE in .env.example
  • Enrichment kept separate from authoritative data — routes/enrich.js
  • Rate limiting — generalLimiter on /api and a tighter enrichLimiter on /api/enrich, backend/server.js:26-53

Quick start commands are copied verbatim from the root, backend and frontend package.json scripts; the endpoint table matches the routes mounted in server.js. No metric was invented.

Only codeql.yml exists as a workflow, so the build badge is the CodeQL one, labelled as such rather than implying a test suite that isn't there.

Screenshots the owner needs to capture are listed in the README's Screenshots section, each with its target path and a pre-written, commented-out image line. Nothing in the README references an image that does not exist.

Only README.md (and a docs/screenshots/README.md placeholder) is touched, to stay clear of the concurrent .project.json / LICENSE work.

🤖 Generated with Claude Code

The README opened with three paragraphs of prose and never said how to run the
project. Add a one-line description, CodeQL/license/language badges, an explicit
note that there is no hosted demo because the dataset is imported locally, and a
Quick start taken verbatim from the scripts in the root, backend and frontend
package.json files, plus the API endpoint table.

Adds five bullets on the decisions worth noticing — the importer that normalizes
USCIS schema drift across fiscal years, storing source-level rows so legitimate
duplicate employer-year records survive, refresh as a first-class operation with
recorded sync metadata, keeping AI enrichment separate from authoritative data,
and rate limiting the enrichment route that costs money per call.

Screenshots cannot be captured here, so that section names the three shots that
would help most with their paths and pre-written, commented-out image lines.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
@bibeshpyakurel
bibeshpyakurel merged commit 1ec133e into main Sep 21, 2026
2 checks passed
@bibeshpyakurel
bibeshpyakurel deleted the docs/readme-refresh branch September 21, 2026 13:40
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant