# Chart generations and edits

Plot publication-ready charts from your data, then change them with follow-up instructions.

Source: https://scifig.ai/docs/api/charts

## Form fields

`POST /charts/generations` takes `multipart/form-data`. Required scope: `charts:generate`.

| Field | Type | Required | Default | Rules |
| --- | --- | --- | --- | --- |
| data | file | yes | — | CSV, TSV, TXT, XLSX or XLS, up to 8 MB, with a header row; the delimiter of CSV, TSV and TXT (comma, tab or semicolon) is detected |
| request | string | yes | — | What to plot, in plain language; 1–16,000 characters |
| chart_types | string | no | chosen from the data | Comma-separated or a JSON array; see the list below |
| journal_style | string | no | `default` | `default`, `nature`, `science`, `cell`, `plos` or `acs` |
| figure_width | string | no | `double` | `single`, `onehalf` or `double` column; PLOS and ACS have no `onehalf` |
| color_palette | string | no | `okabe_ito` | `okabe_ito`, `wong`, `tol_bright`, `tol_muted`, `nature`, `science`, `nejm`, `lancet` or `jama` |
| sheet_name | string | for workbooks with several sheets | — | The worksheet to plot from an XLSX or XLS file (1–255 characters) |

## Data errors

- The data is parsed before anything is created, with the same parser as the website. Data that cannot be read returns `400 INVALID_REQUEST` with `details.field = data` and the parser’s code in `details.reason` (for example `DATACHART_CSV_INVALID`, `DATACHART_TOO_MANY_ROWS`); no job runs and no credits are used.
- A workbook with several sheets and no `sheet_name` returns `400 INVALID_REQUEST` (`DATACHART_WORKSHEET_REQUIRED`) with the sheet names in `details.sheet_names`; send the request again with `sheet_name`. An unknown `sheet_name` returns `DATACHART_WORKSHEET_INVALID`, also with `details.sheet_names`.
- A file type outside the list returns `UNSUPPORTED_DATA_FORMAT`; a file over the size limit returns `DATA_FILE_TOO_LARGE`.
- Each new data file counts against the website’s upload limits; beyond them the answer is `429 RATE_LIMITED` (`UPLOAD_RATE_LIMITED` or `UPLOAD_DAILY_LIMIT_REACHED`). Retrying the same key does not count the file again.

## Chart types

`bar`, `line`, `scatter`, `dot`, `frequency_dot`, `pie`, `nested_pie`, `timeseries`, `waterfall`, `stacked_area`, `radar`, `boxplot`, `violin`, `histogram`, `density`, `raincloud`, `beeswarm`, `qq`, `ogive`, `ecdf`, `ridgeline`, `regression`, `errorbar`, `correlation`, `bland_altman`, `funnel`, `standard_curve`, `bubble`, `hexbin`, `paired_dot`, `volcano`, `kaplan_meier`, `forest`, `roc`, `manhattan`, `ma_plot`, `enrichment`, `dose_response`, `growth_curve`, `enzyme_kinetics`, `precision_recall`, `calibration_curve`, `confusion_matrix`, `heatmap`, `dendrogram`, `upset`, `sankey`, `stem_leaf`, `calendar_heatmap`

## Output and credits

100 credits per chart. A succeeded job has PNG, PDF, TIFF and EPS outputs of the same figure, sized for the journal column. Each output carries `editor_url`, which opens the chart in its DataChart conversation on the website.

## Edits

`POST /charts/edits` with JSON `{"job": "<chart job id>", "instruction": "..."}` changes the chart that job produced, keeping its data, chart types, journal style and palette. `instruction` is 1–16,000 characters. 100 credits, as in the workspace.

- The job must be one of your chart jobs, otherwise `400 INVALID_REQUEST` (`JOB_NOT_FOUND`). If it failed, the answer is `400 INVALID_REQUEST` (`JOB_NOT_SUCCEEDED`).
- If it is still running, the answer is `409 REVISION_CONFLICT` with `details.reason = PARENT_JOB_RUNNING` and `Retry-After`. This one is retryable: send the same request with the same key once the job has finished.
- Only the latest version can be edited, as in the workspace. Edit the newest job of the chain; an older one returns `409 REVISION_CONFLICT` (`NOT_LATEST_REVISION`), also when a newer version was made on the website or another edit of the same version was accepted first.
- One edit at a time: while an edit of the current version is still running, another returns `409 REVISION_CONFLICT` (`EDIT_IN_PROGRESS`). Wait for the running job, then edit its result.
- A chart made on the website before the current chart format cannot be continued through the API: `400 INVALID_REQUEST` (`JOB_NOT_EDITABLE`).
- Edits made through the API appear in the same DataChart conversation on the website.

## cURL

```bash
curl -X POST https://scifig.ai/api/v1/charts/generations \
  -H "Authorization: Bearer $SCIFIG_API_KEY" \
  -H "Idempotency-Key: $(uuidgen)" \
  -H "Prefer: wait=25" \
  -F "data=@measurements.csv" \
  -F "request=Box plot of response by group with individual points" \
  -F "journal_style=nature"
```

## JavaScript

```javascript
import { readFile } from 'node:fs/promises';

const form = new FormData();
form.set('data', new Blob([await readFile('measurements.csv')]), 'measurements.csv');
form.set('request', 'Box plot of response by group with individual points');
form.set('journal_style', 'nature');

const response = await fetch('https://scifig.ai/api/v1/charts/generations', {
  method: 'POST',
  headers: {
    Authorization: `Bearer ${process.env.SCIFIG_API_KEY}`,
    'Idempotency-Key': crypto.randomUUID(),
    Prefer: 'wait=25',
  },
  body: form,
});
const result = await response.json();
console.log(response.status, result);
```

## Python

```python
import os
import uuid

import requests

headers = {
    "Authorization": f"Bearer {os.environ['SCIFIG_API_KEY']}",
    "Idempotency-Key": str(uuid.uuid4()),
    "Prefer": "wait=25",
}
response = requests.post(
    "https://scifig.ai/api/v1/charts/generations",
    headers=headers,
    files={"data": open("measurements.csv", "rb")},
    data={"request": "Box plot of response by group with individual points", "journal_style": "nature"},
    timeout=60,
)
print(response.status_code, response.json())
```