@melaya/runner 1.0.32 → 1.0.34

package/dist/pythonEnv.js

@@ -11,6 +11,7 @@
  */
 import { spawn } from "child_process";
 import { existsSync, mkdirSync, readFileSync, writeFileSync } from "fs";
+import { createHash } from "crypto";
 import { join } from "path";
 import { homedir, platform } from "os";
 import chalk from "chalk";
@@ -86,6 +87,19 @@ export function venvPython() {
         ? join(VENV_DIR, "Scripts", "python.exe")
         : join(VENV_DIR, "bin", "python");
 }
+// The marker is keyed on sharedVersion + a hash of PIP_DEPS so changes
+// to either invalidate the venv. Without the PIP_DEPS half, a runner
+// update that adds a dep (e.g. scrapling[fetchers] in 1.0.32) would
+// silently skip the reinstall on any box whose sharedVersion hadn't
+// also changed — operators would update the runner but never see the
+// new deps. Hashing the full deps list catches additions, removals,
+// AND version-pin changes within deps.
+function _pipDepsHash() {
+    return createHash("sha256").update(JSON.stringify(PIP_DEPS)).digest("hex").slice(0, 12);
+}
+function venvMarkerValue(expectedVersion) {
+    return `${expectedVersion}::${_pipDepsHash()}`;
+}
 function venvIsValid(expectedVersion) {
     if (!existsSync(venvPython()))
         return false;
@@ -93,7 +107,7 @@ function venvIsValid(expectedVersion) {
         return false;
     try {
         const cached = readFileSync(VENV_MARK, "utf-8").trim();
-        return cached === expectedVersion;
+        return cached === venvMarkerValue(expectedVersion);
     }
     catch {
         return false;
@@ -161,16 +175,34 @@ export async function ensurePythonEnv(systemPython, expectedVersion, onProgress
     }
     // Scrapling browser binaries — best-effort. The fast-path AsyncFetcher
     // (curl_cffi TLS impersonation) works without this; only StealthyFetcher
-    // (Cloudflare-bypass rescue path) needs Chromium. If `scrapling install`
-    // fails (offline, network blip, browser-download mirror down) we don't
-    // fail the venv — the search just stays on the fast path. Logs are
-    // surfaced so the operator sees what happened.
+    // (Cloudflare-bypass rescue path) needs Chromium. Scrapling has no
+    // __main__.py — invoke its CLI entry point via `from scrapling.cli
+    // import main; main(['install'])` so the call works the same on every
+    // platform without reaching into the venv's Scripts/bin folder.
+    // Non-fatal: if it fails (offline, mirror down, scrapling.cli rename in
+    // a future version) the search just stays on the fast path.
     onProgress("installing scrapling browsers (best-effort, for stealth rescue path)");
-    const scrCode = await runProc(venvPython(), ["-m", "scrapling", "install"], onProgress);
+    const scrCode = await runProc(venvPython(), ["-c",
+        "import sys\n" +
+            "# Try every known CLI shape across scrapling versions, return 0 on first success.\n" +
+            "tried = []\n" +
+            "for mod, attr in (('scrapling.cli','main'), ('scrapling.__main__','main'), ('scrapling.cli','cli')):\n" +
+            "    try:\n" +
+            "        m = __import__(mod, fromlist=[attr])\n" +
+            "        fn = getattr(m, attr)\n" +
+            "        try: fn(['install'])\n" +
+            "        except TypeError: fn()  # zero-arg variants\n" +
+            "        sys.exit(0)\n" +
+            "    except Exception as e:\n" +
+            "        tried.append(f'{mod}.{attr}: {type(e).__name__}: {e}')\n" +
+            "        continue\n" +
+            "print('scrapling install: no working CLI entry — tried:'); [print(' -', t) for t in tried]\n" +
+            "sys.exit(2)\n",
+    ], onProgress);
     if (scrCode !== 0) {
         onProgress(`(scrapling install exited ${scrCode} — fast-path fetcher still works; stealth-rescue disabled)`);
     }
-    writeFileSync(VENV_MARK, expectedVersion, "utf-8");
+    writeFileSync(VENV_MARK, venvMarkerValue(expectedVersion), "utf-8");
     // sanity: confirm shortuuid + agentscope (via PYTHONPATH) resolve now.
     // Probe must mirror the spawn env so PYTHONPATH=CACHE_DIR points at
     // the cached agentscope source — without this the probe ImportError's

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@melaya/runner",
-  "version": "1.0.32",
+  "version": "1.0.34",
   "description": "Run Melaya AI pipelines locally with your own LM Studio or Ollama models",
   "license": "UNLICENSED",
   "private": false,