External/SnapOtter
1
0
Fork 0
mirror of https://github.com/snapotter-hq/SnapOtter.git synced 2026-08-03 07:46:42 +02:00
Code Issues Packages Projects Releases Wiki Activity

feat(tools): 2.0 phase 5 wave 2 - pdf depth (21 tools) (#220)

Browse Source
This commit is contained in:
SnapOtter
2026-06-13 10:18:55 +08:00
parent ae1337901d
commit 2f39e38162
128 changed files with 11381 additions and 778 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/integration/docs-dispatcher.test.ts
+44 -35
View File
@@ -1,51 +1,33 @@
import { spawnSync } from "node:child_process";
import { existsSync } from "node:fs";
import { existsSync, mkdtempSync, rmSync } from "node:fs";
import { tmpdir } from "node:os";
import { join } from "node:path";
import { getDocsDispatcher, runDocsScript, shutdownDocsDispatcher } from "@snapotter/ai";
import { pdfPageCountPy } from "@snapotter/doc-engine";
import {
pdfMetadataGetPy,
pdfMetadataSetPy,
pdfPageCountPy,
pdfTextPy,
} from "@snapotter/doc-engine";
import { afterAll, describe, expect, it } from "vitest";
/**
* Resolve a Python 3 binary path that actually exists.
* Checks the configured venv first, then falls back to system python3.
* Returns null when no usable python3 is found.
*/
function resolvePython(): string | null {
const venv = process.env.PYTHON_VENV_PATH || join(process.cwd(), ".venv");
if (existsSync(`${venv}/bin/python3`)) return `${venv}/bin/python3`;
// Fall back: find system python3 and derive a "venv" path the bridge accepts
const res = spawnSync("which", ["python3"], { encoding: "utf8" });
if (res.status === 0 && res.stdout.trim()) {
const bin = res.stdout.trim();
// python3 at /opt/homebrew/bin/python3 -> venv prefix = /opt/homebrew
const parts = bin.split("/");
if (parts.length >= 3) {
const prefix = parts.slice(0, -2).join("/");
if (existsSync(`${prefix}/bin/python3`)) return `${prefix}/bin/python3`;
}
}
return null;
}
function pythonHasPikepdf(pythonBin: string): boolean {
const res = spawnSync(pythonBin, ["-c", "import pikepdf"], { encoding: "utf8" });
return res.status === 0;
}
const pythonBin = resolvePython();
const hasPython = pythonBin !== null;
const hasPikepdf = hasPython && pythonHasPikepdf(pythonBin!);
import { hasFitz, hasPikepdf, hasPython } from "../helpers/python-gate.js";
// Ensure the bridge uses a reachable Python; set PYTHON_VENV_PATH before
// any import triggers a dispatcher spawn.
if (hasPython && !existsSync(join(process.cwd(), ".venv", "bin", "python3"))) {
const parts = pythonBin!.split("/");
process.env.PYTHON_VENV_PATH = parts.slice(0, -2).join("/");
// Resolve the system python3 path to derive the venv prefix.
const res = spawnSync("which", ["python3"], { encoding: "utf8" });
if (res.status === 0 && res.stdout.trim()) {
const parts = res.stdout.trim().split("/");
process.env.PYTHON_VENV_PATH = parts.slice(0, -2).join("/");
}
}
if (!hasPython) console.log("[docs-dispatcher] SKIP: no python3 found (venv or system)");
if (hasPython && !hasPikepdf)
console.log("[docs-dispatcher] pikepdf not available; pikepdf tests will skip");
if (hasPython && !hasFitz)
console.log("[docs-dispatcher] fitz (PyMuPDF) not available; fitz tests will skip");
afterAll(async () => {
await shutdownDocsDispatcher();
@@ -71,4 +53,31 @@ describe.skipIf(!hasPikepdf)("docs dispatcher pikepdf (requires python + pikepdf
const pages = await pdfPageCountPy(join(process.cwd(), "tests/fixtures/test-3page.pdf"));
expect(pages).toBe(3);
});
it("metadata get/set roundtrip", async () => {
const pdf = join(process.cwd(), "tests/fixtures/test-3page.pdf");
const dir = mkdtempSync(join(tmpdir(), "meta-"));
try {
const out = join(dir, "meta.pdf");
await pdfMetadataSetPy(pdf, out, { Title: "Wave2" });
const meta = await pdfMetadataGetPy(out);
expect(meta.Title).toBe("Wave2");
} finally {
rmSync(dir, { recursive: true, force: true });
}
});
});
describe.skipIf(!hasFitz)("docs dispatcher fitz (requires python + PyMuPDF)", () => {
it("extracts text with chars > 0", async () => {
const pdf = join(process.cwd(), "tests/fixtures/test-3page.pdf");
const dir = mkdtempSync(join(tmpdir(), "text-"));
try {
const out = join(dir, "text.txt");
const result = await pdfTextPy(pdf, out);
expect(result.chars).toBeGreaterThan(0);
} finally {
rmSync(dir, { recursive: true, force: true });
}
});
});