Post a PDF, get source-linked data

# 1. Upload a PDF and start parsing it.
#    data_class=deidentified states the file has no patient information.
curl -s -X POST "https://parsedocs.tathya.ai/v1/parse?tier=standard&data_class=deidentified" \
  -H "Authorization: Bearer $PARSEDOCS_KEY" \
  -H "Idempotency-Key: $(uuidgen)" \
  -F "file=@fax.pdf"

# 2. Check progress. Repeat until "status" is "completed".
curl -s "https://parsedocs.tathya.ai/v1/jobs/JOB_ID" \
  -H "Authorization: Bearer $PARSEDOCS_KEY"

# 3. Read the structured result, or download the findings as CSV.
curl -s "https://parsedocs.tathya.ai/v1/jobs/JOB_ID/result" \
  -H "Authorization: Bearer $PARSEDOCS_KEY"
curl -s "https://parsedocs.tathya.ai/v1/jobs/JOB_ID/result.csv" \
  -H "Authorization: Bearer $PARSEDOCS_KEY" -o findings.csv

# No file handy? Run a built-in fictional example instead of step 1.
curl -s -X POST "https://parsedocs.tathya.ai/v1/jobs" \
  -H "Authorization: Bearer $PARSEDOCS_KEY" \
  -H "Content-Type: application/json" \
  -d '{"example_id": "referral-nephrology", "tier": "standard"}'

import os, time, uuid
import requests

BASE = "https://parsedocs.tathya.ai"
HEADERS = {"Authorization": f"Bearer {os.environ['PARSEDOCS_KEY']}"}

# Upload a PDF and start parsing it.
with open("fax.pdf", "rb") as f:
    r = requests.post(
        f"{BASE}/v1/parse",
        params={"tier": "standard", "data_class": "deidentified"},
        headers={**HEADERS, "Idempotency-Key": str(uuid.uuid4())},
        files={"file": ("fax.pdf", f, "application/pdf")},
    )
if r.status_code == 402:      # saved, but your credit does not cover it yet
    raise SystemExit(r.json()["error"]["message"])
r.raise_for_status()
job = r.json()["job"]

# Wait for it to finish. Long documents are processed in the background.
while job["status"] in ("queued", "processing"):
    time.sleep(2)
    job = requests.get(f"{BASE}/v1/jobs/{job['id']}", headers=HEADERS).json()["job"]
if job["status"] != "completed":
    raise SystemExit(job.get("failure", {}).get("message", job["status"]))

# Every finding carries the page and the printed line it came from.
result = requests.get(f"{BASE}/v1/jobs/{job['id']}/result", headers=HEADERS).json()
print(result["overview"]["headline"])
for item in result["items"]:
    where = item["evidence"][0] if item["evidence"] else {}
    print(item["entity"], item["value"]["original"], item["value"]["unit"],
          "| page", where.get("page"), "|", item["status"])
print("charged", job["charge"]["amount"], "credit left", job["charge"].get("credit_remaining"))

// Node 18 or newer (built-in fetch, FormData and Blob).
import { readFile } from "node:fs/promises";
import { randomUUID } from "node:crypto";

const BASE = "https://parsedocs.tathya.ai";
const headers = { Authorization: `Bearer ${process.env.PARSEDOCS_KEY}` };

// Upload a PDF and start parsing it.
const form = new FormData();
form.append("file", new Blob([await readFile("fax.pdf")], { type: "application/pdf" }), "fax.pdf");
const started = await fetch(`${BASE}/v1/parse?tier=standard&data_class=deidentified`, {
  method: "POST",
  headers: { ...headers, "Idempotency-Key": randomUUID() },
  body: form,
});
const body = await started.json();
if (!started.ok) throw new Error(body.error.message);   // 402 means: saved, but credit does not cover it yet
let job = body.job;

// Wait for it to finish. Long documents are processed in the background.
while (["queued", "processing"].includes(job.status)) {
  await new Promise((resolve) => setTimeout(resolve, 2000));
  job = (await (await fetch(`${BASE}/v1/jobs/${job.id}`, { headers })).json()).job;
}
if (job.status !== "completed") throw new Error(job.failure?.message ?? job.status);

// Every finding carries the page and the printed line it came from.
const result = await (await fetch(`${BASE}/v1/jobs/${job.id}/result`, { headers })).json();
console.log(result.overview.headline);
for (const item of result.items) {
  console.log(item.entity, item.value.original, item.value.unit, "| page", item.evidence[0]?.page, "|", item.status);
}
console.log("charged", job.charge.amount, "credit left", job.charge.credit_remaining);