jtcsv-codemod

jscodeshift transforms that migrate existing CSV-handling code to jtcsv.

npx jtcsv-codemod papaparse "src/**/*.{js,ts,tsx}"
npx jtcsv-codemod csvtojson "src/**/*.{js,ts,tsx}"

Available transforms

| Transform | From | What it rewrites | |-------------|---------------|----------------------------------------------------------------------------------| | papaparse | papaparse | imports, Papa.parse, Papa.unparse, option names | | csvtojson | csvtojson | imports, csv().fromString/.fromFile/.fromStream(...).subscribe, option names |

More transforms (csv-parser) are planned — open an issue or PR with a code sample if you'd like one.

papaparse → jtcsv

Imports

- import Papa from 'papaparse';
+ import { csvToJson, jsonToCsv } from 'jtcsv';

- const Papa = require('papaparse');
+ const { csvToJson, jsonToCsv } = require('jtcsv');

Calls and option renames

- Papa.parse(csv, { header: true, dynamicTyping: true })
+ csvToJson(csv, { hasHeaders: true, parseNumbers: true })

- Papa.unparse(rows, { delimiter: ';' })
+ jsonToCsv(rows, { delimiter: ';' })

- const rows = Papa.parse(csv).data;
+ const rows = csvToJson(csv);

| Papa option | jtcsv option | Notes | |-------------------|--------------------|--------------------------------------| | header | hasHeaders | | | dynamicTyping | parseNumbers | | | delimiter | delimiter | unchanged | | newline | newline | unchanged | | quoteChar | quoteChar | unchanged | | escapeChar | escapeChar | unchanged | | comments | comments | unchanged | | skipEmptyLines | skipEmptyLines | unchanged |

Dropped options

These have no jtcsv equivalent and are stripped, with a // TODO(jtcsv-codemod): … comment left at the call site so you can review:

worker, download, fastMode, beforeFirstChunk, transformHeader, preview, encoding, chunk, step, complete, error.

If you used step/complete for streaming, switch to:

import { createCsvToJsonStream } from 'jtcsv';
import { createReadStream } from 'fs';

await createReadStream('big.csv')
  .pipe(createCsvToJsonStream({ hasHeaders: true, parseNumbers: true }))
  .on('data', (row) => { /* per-row logic that used to live in step() */ });

csvtojson → jtcsv

Imports

- import csv from 'csvtojson';
+ import { csvToJson } from 'jtcsv';

- const csv = require('csvtojson');
+ const { csvToJson } = require('jtcsv');

The codemod only imports the jtcsv helpers it actually needs (csvToJson, readCsvAsJson, createCsvToJsonStream, createCsvFileToJsonStream) based on the call sites in each file.

Calls

- const rows = await csv().fromString(s);
+ const rows = await csvToJson(s);

- const rows = await csv().fromFile('a.csv');
+ const rows = await readCsvAsJson('a.csv');

Option renames

- csv({ noheader: true, checkType: true, delimiter: ';' }).fromString(s);
+ csvToJson(s, { hasHeaders: false, parseNumbers: true, parseBooleans: true, delimiter: ';' });

Stream subscribe → pipe

- csv().fromStream(readable).subscribe(rowCb, errCb, endCb);
+ readable
+   .pipe(createCsvToJsonStream())
+   .on('data', rowCb)
+   .on('error', errCb)
+   .on('end', endCb);

Options translation

| csvtojson option | jtcsv option | Notes | |------------------|-------------------------------------------|-----------------------------------------------| | noheader | hasHeaders (value inverted) | noheader: true → hasHeaders: false | | delimiter | delimiter | unchanged | | checkType | parseNumbers + parseBooleans | fan-out — both keys emitted | | trim | trim | unchanged | | includeColumns | — (dropped, TODO comment) | no jtcsv equivalent today | | ignoreColumns | — (dropped, TODO comment) | no jtcsv equivalent today | | output | — (dropped, TODO comment) | no jtcsv equivalent today | | colParser | — (dropped, TODO comment) | no jtcsv equivalent today |

Limitations

• .on('json', cb) and .on('end_parsed', cb) — left in place with a TODO; the closest jtcsv equivalent is .on('data', cb) on createCsvToJsonStream, but the row shape differs slightly so the codemod refuses to guess.
• preFileLine(fn) — left in place with a TODO; jtcsv has no per-line preprocessing hook, so you'll want to preprocess the readable stream manually before piping it through createCsvToJsonStream().

CLI flags

The CLI is a thin wrapper around jscodeshift. Anything after -- is passed straight through:

# dry-run (no writes, prints diff)
npx jtcsv-codemod papaparse src/ -- --dry --print

# only .ts and .tsx files
npx jtcsv-codemod papaparse src/ -- --extensions=ts,tsx

Limitations

• Streaming patterns (Papa.parse(file, { step })) are flagged but not rewritten end-to-end — the manual switch is short, see above.
• Web Workers (Papa.parse(file, { worker: true })) require deciding between jtcsv/web-workers and Node worker_threads; we don't guess.
• Comment placement varies slightly by jscodeshift version; review the diff before committing.

Contributing

Source: transforms/papaparse-to-jtcsv.js. Tests: __tests__/papaparse-to-jtcsv.test.js (jscodeshift's applyTransform helper).

PRs welcome.