atsfill 0.0.5.dev191__tar.gz → 0.0.5.dev193__tar.gz

{atsfill-0.0.5.dev191 → atsfill-0.0.5.dev193}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: atsfill
-Version: 0.0.5.dev191
+Version: 0.0.5.dev193
 Summary: Read, dry-run and submit job application forms on hiring systems: Ashby, Greenhouse, Teamtailor, Welcome to the Jungle.
 Keywords: job-applications,ats,ashby,greenhouse,teamtailor,ai-agents,cli
 Author: shbernal
@@ -82,6 +82,7 @@ It is built for agents with a shell: every command prints one JSON document and
 [![pinpoint run history](https://shbernal.github.io/atsfill/run-history/pinpoint.svg)](https://shbernal.github.io/atsfill/run-history/)
 [![hiringthing run history](https://shbernal.github.io/atsfill/run-history/hiringthing.svg)](https://shbernal.github.io/atsfill/run-history/)
 [![recruiterflow run history](https://shbernal.github.io/atsfill/run-history/recruiterflow.svg)](https://shbernal.github.io/atsfill/run-history/)
+[![werecruit run history](https://shbernal.github.io/atsfill/run-history/werecruit.svg)](https://shbernal.github.io/atsfill/run-history/)
 [![jobaffinity run history](https://shbernal.github.io/atsfill/run-history/jobaffinity.svg)](https://shbernal.github.io/atsfill/run-history/)
 [![welcomekit run history](https://shbernal.github.io/atsfill/run-history/welcomekit.svg)](https://shbernal.github.io/atsfill/run-history/)
 [![join run history](https://shbernal.github.io/atsfill/run-history/join.svg)](https://shbernal.github.io/atsfill/run-history/)

{atsfill-0.0.5.dev191 → atsfill-0.0.5.dev193}/README.md

@@ -53,6 +53,7 @@ It is built for agents with a shell: every command prints one JSON document and
 [![pinpoint run history](https://shbernal.github.io/atsfill/run-history/pinpoint.svg)](https://shbernal.github.io/atsfill/run-history/)
 [![hiringthing run history](https://shbernal.github.io/atsfill/run-history/hiringthing.svg)](https://shbernal.github.io/atsfill/run-history/)
 [![recruiterflow run history](https://shbernal.github.io/atsfill/run-history/recruiterflow.svg)](https://shbernal.github.io/atsfill/run-history/)
+[![werecruit run history](https://shbernal.github.io/atsfill/run-history/werecruit.svg)](https://shbernal.github.io/atsfill/run-history/)
 [![jobaffinity run history](https://shbernal.github.io/atsfill/run-history/jobaffinity.svg)](https://shbernal.github.io/atsfill/run-history/)
 [![welcomekit run history](https://shbernal.github.io/atsfill/run-history/welcomekit.svg)](https://shbernal.github.io/atsfill/run-history/)
 [![join run history](https://shbernal.github.io/atsfill/run-history/join.svg)](https://shbernal.github.io/atsfill/run-history/)

{atsfill-0.0.5.dev191 → atsfill-0.0.5.dev193}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "atsfill"
-version = "0.0.5.dev191"
+version = "0.0.5.dev193"
 description = "Read, dry-run and submit job application forms on hiring systems: Ashby, Greenhouse, Teamtailor, Welcome to the Jungle."
 readme = "README.md"
 license = "MIT"

{atsfill-0.0.5.dev191 → atsfill-0.0.5.dev193}/pyproject.toml.orig

@@ -1,6 +1,6 @@
 [project]
 name = "atsfill"
-version = "0.0.5.dev191"
+version = "0.0.5.dev193"
 description = "Read, dry-run and submit job application forms on hiring systems: Ashby, Greenhouse, Teamtailor, Welcome to the Jungle."
 readme = "README.md"
 license = "MIT"

{atsfill-0.0.5.dev191 → atsfill-0.0.5.dev193}/src/atsfill/adapters/__init__.py

@@ -30,6 +30,7 @@ from atsfill.adapters.successfactors import SuccessFactors
 from atsfill.adapters.teamtailor import Teamtailor
 from atsfill.adapters.ukg import UKG
 from atsfill.adapters.welcomekit import WelcomeKit
+from atsfill.adapters.werecruit import WeRecruit
 from atsfill.adapters.workable import Workable
 from atsfill.adapters.workatastartup import WorkAtAStartup
 from atsfill.adapters.workday import Workday
@@ -49,6 +50,7 @@ def adapters() -> list[Adapter]:
         Join(),
         JobAffinity(),
         WelcomeKit(),
+        WeRecruit(),
         SmartRecruiters(),
         Oracle(),
         Manatal(),

{atsfill-0.0.5.dev191 → atsfill-0.0.5.dev193}/src/atsfill/adapters/phenom.py

@@ -25,7 +25,7 @@ from typing import Any
 from urllib.parse import parse_qs, urlsplit
 
 from atsfill.core import browser, matching, paths
-from atsfill.core.adapter import Gap, NotTaken, Unsupported, known_gap
+from atsfill.core.adapter import Blocked, Gap, NotTaken, Unsupported, known_gap
 from atsfill.core.browser import HTTP_TIMEOUT_S, BrowserError, Locator, Overlaid, Page, Timeout
 from atsfill.core.results import (
     AllOf,
@@ -238,22 +238,24 @@ _FETCH = """async url => {
   return response.json();
 }"""
 
-# Posts each body to `/applySubmit`, `most` at a time; null for one that fails.
+# Posts each body to `/applySubmit`, `most` at a time; null for one that fails. With the
+# answers, whether one was refused (403), after which none is posted.
 _POSTS = """async ([bodies, token, most]) => {
   const out = bodies.map(() => null);
-  let next = 0;
+  let next = 0, refused = false;
   const post = async () => {
-    for (let i = next++; i < bodies.length; i = next++) {
+    for (let i = next++; i < bodies.length && !refused; i = next++) {
       try {
         const response = await fetch('/applySubmit', {method: 'POST',
           body: JSON.stringify(bodies[i]),
           headers: {'Content-Type': 'application/json', 'x-csrf-token': token}});
         if (response.ok) out[i] = await response.json();
+        refused ||= response.status === 403;
       } catch {}
     }
   };
   await Promise.all(Array.from({length: most}, post));
-  return out;
+  return [out, refused];
 }"""
 
 
@@ -350,14 +352,22 @@ class Phenom:
         except ValueError as e:
             raise RuntimeError(str(e)) from None
         token = get(ph, "sessionParams.csrfToken") or ""
+
+        refused = False
+
+        def posts(bodies: list[dict[str, Any]]) -> list[Any]:
+            nonlocal refused
+            out, now = page.evaluate(_POSTS, [bodies, token, LOADERS_AT_ONCE])
+            refused |= now
+            return out
+
         try:
-            fields, raws = _read(
-                ph,
-                lambda url: page.evaluate(_FETCH, url),
-                lambda bodies: page.evaluate(_POSTS, [bodies, token, LOADERS_AT_ONCE]),
-            )
+            fields, raws = _read(ph, lambda url: page.evaluate(_FETCH, url), posts)
         except (BrowserError, ValueError, KeyError) as e:
             raise RuntimeError(f"no configuration for the wizard ({e})") from None
+        finally:  # raised once the read ends, so the answers that came are kept
+            if refused:
+                raise Blocked("rate_limited: the careers site answers /applySubmit 403")
         self._raws = {raw["id"]: raw for raw in raws}
         return fields
 

atsfill-0.0.5.dev193/src/atsfill/adapters/werecruit.py

@@ -0,0 +1,230 @@
+"""WeRecruit (`careers.werecruit.io/<lang>/<company>/offres/<slug>-<id>`), as
+docs/adapters/werecruit.md observes.
+
+The apply form is on the posting page, rendered on the server: `request_read` fetches it over
+HTTP, `browser_read` takes it from the browser, and both list the fields with one parser. The
+documents go with the form's one multipart post, back to the posting's URL.
+"""
+
+import re
+import urllib.request
+from dataclasses import dataclass
+from urllib.error import HTTPError
+from urllib.parse import urlsplit
+
+from atsfill.core import html
+from atsfill.core.adapter import Gap, Unsupported
+from atsfill.core.browser import HTTP_TIMEOUT_S, Page, Timeout
+from atsfill.core.results import Answer, Field, FieldType, SideEffect
+
+HOST = "careers.werecruit.io"
+# A posting on WeRecruit's host names its company; on an employer's own domain it does not.
+HOSTED = re.compile(r"^/(?:fr|en)/([\w-]+)/(?:offres|offers)/(?:[\w-]*-)?([0-9a-f]{6})/?$")
+OWN_DOMAIN = re.compile(r"^/(?:fr|en)/(?:offres|offers)/(?:[\w-]*-)?([0-9a-f]{6})/?$")
+FORM = "form#form"
+PORTAL = f"{FORM} input[name=isEmployerPortal]"  # WeRecruit's form, whatever the domain
+FORM_TIMEOUT_MS = 30_000
+SUBMIT_TIMEOUT_MS = 90_000
+KINDS = {
+    "text": FieldType.TEXT,
+    "email": FieldType.TEXT,
+    "tel": FieldType.PHONE,
+    "textarea": FieldType.TEXT,
+    "select": FieldType.CHOICE,
+    "radio": FieldType.CHOICE,
+    "file": FieldType.FILE,
+}
+
+# What each control holds, by name: a value, a ticked radio's id, a document's file name.
+_READ_BACK = """(form, names) => {
+  const out = {};
+  for (const name of names) {
+    const els = [...form.querySelectorAll(`[name="${name}"]`)];
+    const el = els[0];
+    if (!el) continue;
+    if (el.type === "radio") out[name] = els.find(e => e.checked)?.id ?? "";
+    else if (el.type === "file") out[name] = el.files[0]?.name ?? "";
+    else out[name] = el.value ?? "";
+  }
+  return out;
+}"""
+
+
+@dataclass
+class _Control:
+    field: Field
+    name: str  # the control's own name, which may carry spaces (` photo`)
+    kind: str
+    options: dict[str, str]  # option text by value, or by radio id
+
+
+class WeRecruit:
+    name = "werecruit"
+    known_gaps: dict[str, Gap] = {}
+    side_effects = [SideEffect("submit", "employer", "the application reaches the employer")]
+
+    def __init__(self) -> None:
+        self._controls: list[_Control] = []
+
+    def detect(self, url: str, page: Page | None) -> bool:
+        """A posting on WeRecruit's host, or one on an employer's domain that serves its form."""
+        parts = urlsplit(url)
+        if parts.hostname == HOST:
+            return HOSTED.match(parts.path) is not None
+        if not OWN_DOMAIN.match(parts.path):
+            return False
+        return page is not None and page.locator(PORTAL).count() > 0
+
+    def posting(self, url: str) -> str:
+        """The company and the id: the slug and the language do not change the posting."""
+        parts = urlsplit(url)
+        if parts.hostname == HOST and (m := HOSTED.match(parts.path)):
+            return f"https://{HOST}/fr/{m[1]}/offres/{m[2]}"
+        if m := OWN_DOMAIN.match(parts.path):
+            return f"https://{parts.hostname}/fr/offres/{m[1]}"
+        return url
+
+    def request_read(self, url: str) -> list[Field] | None:
+        if not self.detect(url, None):
+            return None
+        try:
+            controls = _parse(_get(url.split("?")[0]))
+        except HTTPError, OSError, ValueError:
+            return None
+        return [c.field for c in controls] if controls is not None else None
+
+    def browser_read(self, page: Page) -> list[Field]:
+        try:
+            page.locator(FORM).wait_for(state="attached", timeout=FORM_TIMEOUT_MS)
+        except Timeout:
+            raise Unsupported("posting_gone: no application form on the page") from None
+        controls = _parse(page.content())
+        assert controls is not None
+        self._controls = controls
+        return [c.field for c in controls]
+
+    def fill(self, page: Page, answers: dict[str, Answer]) -> None:
+        form = page.locator(FORM)
+        for c in self._controls:
+            if c.field.id not in answers or c.field.unfillable:
+                continue
+            value = _one(answers[c.field.id])
+            control = form.locator(f'[name="{c.name}"]')
+            match c.kind:
+                case "select":
+                    by_text = {t: v for v, t in c.options.items()}
+                    control.select_option(value=by_text.get(value, value))
+                case "radio":
+                    radio = next((i for i, t in c.options.items() if t == value), value)
+                    form.locator(f'label[for="{radio}"]').click()
+                case "file":
+                    control.set_input_files(value)
+                case _:
+                    control.fill(value)
+
+    def read_back(self, page: Page) -> dict[str, Answer]:
+        shown = [c for c in self._controls if not c.field.unfillable]
+        values = page.locator(FORM).evaluate(_READ_BACK, [c.name for c in shown])
+        out: dict[str, Answer] = {}
+        for c in shown:
+            value = values.get(c.name)
+            if c.kind in ("select", "radio"):
+                value = c.options.get(value, value)
+            if value not in (None, ""):
+                out[c.field.id] = value
+        return out
+
+    def submit(self, page: Page) -> str | None:
+        """Clicks Je postule; the evidence is the URL and status of the form's post. The page
+        validates first: no post, and its errors are the failure."""
+        action = page.locator(FORM).get_attribute("action") or page.url
+        try:
+            with page.expect_response(
+                lambda r: r.request.method == "POST" and r.url.split("?")[0] == action,
+                timeout=SUBMIT_TIMEOUT_MS,
+            ) as posted:
+                page.locator("#submit-button").click()
+        except Timeout:
+            raise RuntimeError(f"the form was not posted: {_errors(page)}") from None
+        if posted.value.status >= 400:
+            raise RuntimeError(f"the post answered {posted.value.status}")
+        return f"{posted.value.url} {posted.value.status}"
+
+
+def _get(url: str) -> str:
+    request = urllib.request.Request(url, headers={"User-Agent": "atsfill"})
+    with urllib.request.urlopen(request, timeout=HTTP_TIMEOUT_S) as response:
+        return response.read().decode()
+
+
+def _parse(source: str) -> list[_Control] | None:
+    """Each named control of the apply form, a radio group once; None without the form."""
+    nodes = html.parse(source).walk()
+    form = next((n for n in nodes if n.tag == "form" and n.attrs.get("id") == "form"), None)
+    if form is None:
+        return None
+    labels = {n.attrs["for"]: n for n in form.walk() if n.tag == "label" and "for" in n.attrs}
+    out: list[_Control] = []
+    seen: set[str] = set()
+    for node in form.walk():
+        name = node.attrs.get("name")
+        if node.tag not in ("input", "select", "textarea") or not name or name in seen:
+            continue
+        kind = node.attrs.get("type", "text") if node.tag == "input" else node.tag
+        if kind in ("hidden", "submit", "button"):
+            continue
+        seen.add(name)
+        group = node.closest(lambda n: "form-group" in n.classes())
+        own = group.find(lambda n: n.tag == "label") if group else None
+        label = _label(own) if own is not None else name.strip()
+        required = "required" in node.attrs
+        fid = name.strip()
+        if kind not in KINDS:
+            reason = f"unhandled_rendering: {fid} is a {kind}"
+            field = Field(fid, label, FieldType.TEXT, required, unfillable=reason)
+            out.append(_Control(field, name, kind, {}))
+            continue
+        options: dict[str, str] = {}
+        if kind == "select":
+            options = {
+                o.attrs["value"]: " ".join(o.text().split())
+                for o in node.walk()
+                if o.tag == "option" and o.attrs.get("value")
+            }
+        elif kind == "radio":
+            options = {
+                r.attrs["id"]: _label(labels[r.attrs["id"]])
+                for r in form.walk()
+                if r.tag == "input" and r.attrs.get("name") == name and r.attrs.get("id") in labels
+            }
+        accept = node.attrs.get("accept", "").split(",") if kind == "file" else []
+        field = Field(
+            fid,
+            label,
+            KINDS[kind],
+            required,
+            list(options.values()) or None,
+            accept=[a.strip() for a in accept if a.strip()] or None,
+        )
+        out.append(_Control(field, name, kind, options))
+    return out
+
+
+def _label(node: html.Node) -> str:
+    """A label's text, without its required marker."""
+
+    def text(n: html.Node | str) -> str:
+        if isinstance(n, str):
+            return n
+        return "" if "required" in n.classes() else "".join(text(c) for c in n.children)
+
+    return " ".join(text(node).split())
+
+
+def _one(answer: Answer) -> str:
+    return answer if isinstance(answer, str) else answer[0]
+
+
+def _errors(page: Page) -> str:
+    shown = page.locator(f"{FORM} .invalid-feedback:visible, #errorMessageDiv:visible")
+    return " ".join(" ".join(shown.all_inner_texts()).split())[:200] or "no error shown"

{atsfill-0.0.5.dev191 → atsfill-0.0.5.dev193}/src/atsfill/registry.toml

@@ -429,3 +429,11 @@ account = "none"
 hosting = ["hosted"]
 adapter = "welcomekit"
 scope = { is = "supported", date = 2026-10-29, commit = "238798d" }
+
+[werecruit]
+kind = "ats"
+share = 0.0007
+account = "none"
+hosting = ["hosted"]
+adapter = "werecruit"
+scope = { is = "supported", date = 2026-10-29, commit = "cf7634f" }