fix: auto-detect jobspy venv so contributors don't need PYTHON_PATH (#293)

CONTRIBUTING.md

@@ -34,6 +34,15 @@ npm --workspace orchestrator run db:migrate
 npm --workspace orchestrator run dev
 ```
 
+If you are working with extractors that use Glassdoor, Indeed, or LinkedIn (powered by python-jobspy), set up the Python venv once:
+
+```bash
+python3 -m venv extractors/jobspy/.venv
+extractors/jobspy/.venv/bin/pip install -r extractors/jobspy/requirements.txt
+```
+
+The runner auto-detects the venv — no need to set `PYTHON_PATH`.
+
 If you are editing docs:
 
 ```bash

extractors/jobspy/src/run.ts

@@ -1,4 +1,5 @@
 import { spawn } from "node:child_process";
+import { existsSync } from "node:fs";
 import { mkdir, readFile, unlink } from "node:fs/promises";
 import { dirname, join } from "node:path";
 import { createInterface } from "node:readline";
@@ -180,8 +181,19 @@ export async function runJobSpy(
         const outputJson = join(OUTPUT_DIR, `jobspy_jobs_${suffix}.json`);
 
         await new Promise<void>((resolve, reject) => {
+          // Auto-detect venv if present, so contributors don't need to set
+          // PYTHON_PATH manually. The venv is created once with:
+          //   python3 -m venv .venv && .venv/bin/pip install -r requirements.txt
+          // In Docker, PYTHON_PATH is set explicitly to /usr/bin/python3.
+          const venvPython = join(
+            EXTRACTOR_DIR,
+            ".venv",
+            process.platform === "win32" ? "Scripts/python.exe" : "bin/python3",
+          );
           const pythonPath = process.env.PYTHON_PATH
             ? process.env.PYTHON_PATH
+            : existsSync(venvPython)
+              ? venvPython
               : process.platform === "win32"
                 ? "python"
                 : "python3";