atsfill 0.0.5.dev181__tar.gz → 0.0.5.dev184__tar.gz

{atsfill-0.0.5.dev181 → atsfill-0.0.5.dev184}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: atsfill
-Version: 0.0.5.dev181
+Version: 0.0.5.dev184
 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

{atsfill-0.0.5.dev181 → atsfill-0.0.5.dev184}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "atsfill"
-version = "0.0.5.dev181"
+version = "0.0.5.dev184"
 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.dev181 → atsfill-0.0.5.dev184}/pyproject.toml.orig

@@ -1,6 +1,6 @@
 [project]
 name = "atsfill"
-version = "0.0.5.dev181"
+version = "0.0.5.dev184"
 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.dev181 → atsfill-0.0.5.dev184}/src/atsfill/adapters/ashby.py

@@ -78,7 +78,9 @@ class Ashby:
 
     def __init__(self) -> None:
         self._forms: list[dict[str, Any]] = []  # the application form, then each survey
-        self._pending: set[Any] = set()  # GraphQL requests in flight, since `browser_read`
+        self._pending: set[Any] = set()  # GraphQL requests in flight, since `open_form`
+        self._listening = False
+        self._loaded: Response | None = None  # the ApiJobPosting response `open_form` caught
         self._last_change = 0.0
         self._hosted: str | None = None  # the jobs.ashbyhq.com posting a custom domain embeds
 
@@ -125,14 +127,24 @@ class Ashby:
         except OSError, ValueError, KeyError, TypeError, Unsupported:
             return None
 
-    def browser_read(self, page: Page) -> list[Field]:
-        page.on("request", self._on_request)
-        page.on("requestfinished", self._on_finished)
-        page.on("requestfailed", self._on_finished)
-        page.on("response", self._on_response)
+    def open_form(self, page: Page, url: str) -> None:
+        """Loads `/application`, listening first for the page's own ApiJobPosting response."""
+        if not self._listening:
+            page.on("request", self._on_request)
+            page.on("requestfinished", self._on_finished)
+            page.on("requestfailed", self._on_finished)
+            page.on("response", self._on_response)
+            self._listening = True
         with page.expect_response(lambda r: _op(r) == "ApiJobPosting") as posting:
-            page.goto(self.posting(page.url) + "/application")
-        job = posting.value.json()["data"]["jobPosting"]
+            page.goto(self.posting(url) + "/application")
+        self._loaded = posting.value
+
+    def browser_read(self, page: Page) -> list[Field]:
+        if self._loaded is None:  # a page claimed once loaded: an embed
+            self.open_form(page, page.url)
+        assert self._loaded
+        loaded, self._loaded = self._loaded, None
+        job = loaded.json()["data"]["jobPosting"]
         if job is None:
             raise Unsupported("posting_gone")
         self._forms = [job["applicationForm"], *job["surveyForms"]]

{atsfill-0.0.5.dev181 → atsfill-0.0.5.dev184}/src/atsfill/adapters/breezy.py

@@ -211,6 +211,10 @@ _FILE_SHOWN = """e => {
 }"""
 
 
+def _on_form(url: str) -> bool:
+    return urlsplit(url).path.rstrip("/").endswith("/apply")
+
+
 class Breezy:
     name = "breezy"
     known_gaps: dict[str, Gap] = {}
@@ -236,9 +240,12 @@ class Breezy:
         match = JOB.match(parts.path)
         return f"https://{parts.hostname}/p/{match[1]}" if match else url
 
+    def open_form(self, page: Page, url: str) -> None:
+        page.goto(url if _on_form(url) else f"{self.posting(url)}/apply")
+
     def browser_read(self, page: Page) -> list[Field]:
-        if not urlsplit(page.url).path.rstrip("/").endswith("/apply"):
-            page.goto(f"{self.posting(page.url)}/apply")
+        if not _on_form(page.url):
+            self.open_form(page, page.url)
         form = page.locator(FORM)
         try:
             form.locator("[name=cName]").wait_for(state="attached", timeout=FORM_TIMEOUT_MS)

{atsfill-0.0.5.dev181 → atsfill-0.0.5.dev184}/src/atsfill/adapters/jobstreet.py

@@ -34,6 +34,8 @@ APPLY = "[data-automation=job-detail-apply]"
 CONTINUE = "[data-testid=continue-button]"
 SUBMIT = "[data-testid=review-submit-application]"
 SAVE_DETAILS = "[data-automation=personal-details-save-button]"
+RESUME_FILE = "input[type=file][id^=resume]"
+LETTER_FILE = "input[type=file][id^=coverLetter]"
 CHALLENGE = "Just a moment..."  # Cloudflare's interstitial: it passes on its own, or on a reload
 CHALLENGE_TRIES = 2
 SIGN_IN_TIMEOUT_MS = 90_000  # SEEK's sign-in takes up to ~50 s to show its form
@@ -53,6 +55,7 @@ SUGGEST_TIMEOUT_S = 8
 POLL_MS = 250
 KEY_DELAY_MS = 40
 DETAILS = ("firstName", "lastName", "currentLocation", "phone")
+DOCUMENTS = ("resume", "coverLetter", "coverLetterFile")
 VISIBILITY = ["Public", "Standard", "Limited", "Hidden"]
 
 _STEP_ONE = """() => {
@@ -62,7 +65,7 @@ _STEP_ONE = """() => {
     .find(e => e.checked);
   const label = e => e && document.querySelector(`label[for="${e.id}"]`)?.innerText.trim();
   const resume = picked('document-select');
-  const letter = document.querySelector('[data-testid=coverLetter-method-change]');
+  const method = picked('coverLetter-method');
   const phone = value('#phoneNumber');
   return {
     firstName: value('#firstName'), lastName: value('#lastName'),
@@ -70,7 +73,8 @@ _STEP_ONE = """() => {
     phone: phone && code?.value
       ? `${code.options[code.selectedIndex].text.match(/\\+\\d+/)?.[0] ?? ''}${phone}` : null,
     resume: resume && resume.value !== 'dont-include' ? label(resume) : null,
-    coverLetter: letter?.checked ? value('[data-testid=coverLetterTextInput]') : null,
+    coverLetter: method?.value === 'change' ? value('[data-testid=coverLetterTextInput]') : null,
+    coverLetterFile: method?.value === 'upload' ? label(method) : null,
   };
 }"""
 
@@ -198,7 +202,7 @@ class Jobstreet:
         self._read |= _personal_details(page, self._fields, answers)
         _documents_fill(page, answers)
         documents = page.evaluate(_STEP_ONE)
-        self._read |= {k: documents[k] for k in ("resume", "coverLetter") if documents[k]}
+        self._read |= {k: documents[k] for k in DOCUMENTS if documents[k]}
         _continue(page)
         if page.url.split("?")[0].endswith("/role-requirements"):
             for field_id, answer in answers.items():
@@ -316,14 +320,12 @@ def _operations(response: Response) -> dict[str, Any]:
 
 def _documents(page: Page, viewer: dict[str, Any]) -> list[Field]:
     """The first step: the profile's personal details, then the resume and the cover letter,
-    optional where the page offers to leave them out."""
+    optional where the page offers to leave them out. The cover letter is written or uploaded,
+    two alternatives."""
     details = viewer["personalDetails"]
     code = (details.get("countryCallingCode") or {}).get("countryCallingCode", "")
     location = (viewer.get("currentLocation2") or {}).get("description")
-    resume = page.locator("input[type=file][id^=resume]")
-    accept: list[str] | None = None
-    if resume.count():
-        accept = [str(a) for a in (resume.get_attribute("accept") or "").split(",")]
+    letter = not page.locator("[data-testid=coverLetter-method-none]").count()
     return [
         Field("firstName", "First name", FieldType.TEXT, True,
               prefilled=details.get("firstName") or None),
@@ -333,12 +335,21 @@ def _documents(page: Page, viewer: dict[str, Any]) -> list[Field]:
         Field("phone", "Phone number", FieldType.PHONE, False,
               prefilled=f"{code}{details['phoneNumber']}" if details.get("phoneNumber") else None),
         Field("resume", "Resumé", FieldType.FILE,
-              not page.locator("input[type=radio][value=dont-include]").count(), accept=accept),
-        Field("coverLetter", "Cover letter", FieldType.TEXT,
-              not page.locator("[data-testid=coverLetter-method-none]").count()),
+              not page.locator("input[type=radio][value=dont-include]").count(),
+              accept=_accept(page, RESUME_FILE)),
+        Field("coverLetter", "Cover letter", FieldType.TEXT, letter, one_of="coverLetter"),
+        Field("coverLetterFile", "Cover letter", FieldType.FILE, letter,
+              accept=_accept(page, LETTER_FILE), one_of="coverLetter"),
     ]  # fmt: skip
 
 
+def _accept(page: Page, selector: str) -> list[str] | None:
+    upload = page.locator(selector)
+    if not upload.count():
+        return None
+    return [str(a) for a in (upload.get_attribute("accept") or "").split(",")]
+
+
 def _question(raw: dict[str, Any]) -> Field | None:
     field_id, label = f"question.{raw['id']}", raw["text"]
     options = [o["text"] for o in raw.get("options") or []]
@@ -404,20 +415,27 @@ def _phone(page: Page, answer: str) -> None:
 
 def _documents_fill(page: Page, answers: dict[str, Answer]) -> None:
     """Uploads the resume, which the page then picks, or leaves it out; the same for the cover
-    letter, written in its box."""
+    letter, uploaded or written in its box."""
     if (resume := answers.get("resume")) is not None:
-        page.locator("input[type=file][id^=resume]").set_input_files(str(resume))
-        name = str(resume).rsplit("/", 1)[-1]
-        page.locator("label", has_text=name).first.wait_for(timeout=SAVE_TIMEOUT_MS)
+        _upload(page, RESUME_FILE, str(resume))
     elif (none := page.locator("input[type=radio][value=dont-include]")).count():
         none.check(force=True)
-    if (letter := answers.get("coverLetter")) is not None:
-        page.locator("[data-testid=coverLetter-method-change]").check(force=True)
+    if (letter := answers.get("coverLetterFile")) is not None:
+        _upload(page, LETTER_FILE, str(letter))
+    elif (letter := answers.get("coverLetter")) is not None:
+        page.locator("input[type=radio][data-testid=coverLetter-method-change]").check(force=True)
         page.locator("[data-testid=coverLetterTextInput]").fill(str(letter))
     elif (none := page.locator("[data-testid=coverLetter-method-none]")).count():
         none.check(force=True)
 
 
+def _upload(page: Page, selector: str, path: str) -> None:
+    """Sets the file, then waits for the radio the page adds and picks, labelled with its name."""
+    page.locator(selector).set_input_files(path)
+    name = path.rsplit("/", 1)[-1]
+    page.locator("label", has_text=name).first.wait_for(timeout=SAVE_TIMEOUT_MS)
+
+
 def _answer(page: Page, field: Field, answer: Answer) -> None:
     name = f"questionnaire.{field.id.removeprefix('question.')}"
     select = page.locator(f'select[name="{name}"]')

{atsfill-0.0.5.dev181 → atsfill-0.0.5.dev184}/src/atsfill/adapters/lever.py

@@ -116,6 +116,10 @@ _READ_BACK = """(form, kinds) => {
 }"""
 
 
+def _on_form(url: str) -> bool:
+    return urlsplit(url).path.rstrip("/").endswith("/apply")
+
+
 class Lever:
     name = "lever"
     known_gaps: dict[str, Gap] = {}
@@ -134,9 +138,12 @@ class Lever:
         parts, job = urlsplit(url), _job(url)
         return f"https://{parts.hostname}/{job[0]}/{job[1]}" if job else url
 
+    def open_form(self, page: Page, url: str) -> None:
+        page.goto(url if _on_form(url) else f"{self.posting(url)}/apply")
+
     def browser_read(self, page: Page) -> list[Field]:
-        if not urlsplit(page.url).path.rstrip("/").endswith("/apply"):
-            page.goto(f"{self.posting(page.url)}/apply")
+        if not _on_form(page.url):
+            self.open_form(page, page.url)
         form = page.locator(FORM)
         try:
             form.wait_for(state="attached", timeout=FORM_TIMEOUT_MS)

{atsfill-0.0.5.dev181 → atsfill-0.0.5.dev184}/src/atsfill/adapters/manatal.py

@@ -12,7 +12,7 @@ from typing import Any
 from urllib.parse import urlsplit
 
 from atsfill.core.adapter import Blocked, Gap, Unsupported, known_gap
-from atsfill.core.browser import HTTP_TIMEOUT_S, Page
+from atsfill.core.browser import HTTP_TIMEOUT_S, Page, Response
 from atsfill.core.results import Answer, Field, FieldType, Is, SideEffect
 
 HOST = "www.careers-page.com"
@@ -183,6 +183,7 @@ class Manatal:
 
     def __init__(self) -> None:
         self._entries: list[dict[str, Any]] = []
+        self._loaded: Response | None = None  # the form's GET `open_form` caught
 
     def detect(self, url: str, page: Page | None) -> bool:
         parts = urlsplit(url).path.strip("/").split("/")
@@ -210,13 +211,20 @@ class Manatal:
         except OSError, ValueError, KeyError, TypeError:
             return None
 
-    def browser_read(self, page: Page) -> list[Field]:
+    def open_form(self, page: Page, url: str) -> None:
+        """Loads `/apply`, listening first for the form's GET."""
         with page.expect_response(
             lambda r: r.request.method == "GET" and "/application-form/" in r.url,
             timeout=LOAD_TIMEOUT_MS,
         ) as loaded:
-            page.goto(_posting(page.url) + "/apply")
-        response = loaded.value
+            page.goto(_posting(url) + "/apply")
+        self._loaded = loaded.value
+
+    def browser_read(self, page: Page) -> list[Field]:
+        if self._loaded is None:
+            self.open_form(page, page.url)
+        assert self._loaded
+        response, self._loaded = self._loaded, None
         if response.headers.get("x-amzn-waf-action") == "challenge":
             raise Blocked("captcha: AWS WAF challenge")
         if not response.ok:

{atsfill-0.0.5.dev181 → atsfill-0.0.5.dev184}/src/atsfill/adapters/phenom.py

@@ -44,6 +44,7 @@ SEQ = re.compile(r"^/[a-z]{2,8}/[a-z]{2}/job/([A-Za-z0-9]+)")
 CDN = "https://cdn-prod-static.phenompeople.com/CareerConnectResources"
 FORM = "form.rjsf"
 NEXT = "button#next"
+TERMS = "terms"  # a notice some sites' wizard opens on, its one button accepting it
 SUBMIT = "button#submit, button.btn-submit, button[type=submit].btn-navigate:not(#next)"
 RESUME = ".resume-upload-wrapper input[type=file]"
 RESUME_ID = "resume"
@@ -264,6 +265,11 @@ class Phenom:
         SideEffect(
             "upload", "hiring_system", "the resume is uploaded to Phenom and parsed at once"
         ),
+        SideEffect(
+            "fill",
+            "hiring_system",
+            "an employer notice the wizard opens on (`page=terms`) is accepted to reach the form",
+        ),
         SideEffect(
             "fill",
             "hiring_system",
@@ -355,6 +361,7 @@ class Phenom:
     def fill(self, page: Page, answers: dict[str, Answer]) -> None:
         page.context.clear_cookies()
         self._open(page, fresh=True)
+        self._accept_terms(page)
         self._kept, self._shown = {}, {}
         if RESUME_ID in answers and page.locator(RESUME).count():
             _upload(page, _one(answers[RESUME_ID]))
@@ -414,17 +421,31 @@ class Phenom:
         if fresh or not page.url.startswith(self._apply):
             page.goto(self._apply, wait_until="domcontentloaded")
         _check_open(page)
+        self._wait_for_step(page)
+
+    def _wait_for_step(self, page: Page) -> None:
         try:
             page.locator(f"{FORM} {NEXT}, {FORM} button[type=submit]").first.wait_for(
                 state="attached", timeout=FORM_TIMEOUT_MS
             )
         except Timeout:
-            query = parse_qs(urlsplit(page.url).query)
-            if step := (query.get("stepname") or query.get("page") or [None])[0]:
+            if step := _step_name(page.url):
                 gap = known_gap(self, "first_step", f"the wizard opens on {step!r}")
                 raise Unsupported(gap) from None
             raise RuntimeError("no application step on the page") from None
 
+    def _accept_terms(self, page: Page) -> None:
+        """Accepts the terms some sites' wizard opens on, a step with no field before the form."""
+        if _step_name(page.url) != TERMS:
+            return
+        _click(page, page.locator(FORM).locator(SUBMIT).first)
+        try:
+            page.wait_for_url(lambda url: _step_name(url) != TERMS, timeout=STEP_TIMEOUT_MS)
+        except Timeout:
+            gap = known_gap(self, "first_step", "the wizard stays on its terms once accepted")
+            raise Unsupported(gap) from None
+        self._wait_for_step(page)
+
     def _step(self, page: Page, answers: dict[str, Answer]) -> list[dict[str, Any]]:
         """The answered fields the current step shows, in the configuration's order."""
         shown = {r["id"]: r for r in page.locator(FORM).evaluate(_SCAN)}
@@ -1237,6 +1258,12 @@ def _seq(url: str) -> str | None:
     return parse_qs(parts.query).get("jobSeqNo", [None])[0]
 
 
+def _step_name(url: str) -> str | None:
+    """The wizard's step as its URL names it (`stepname=`, `page=`), if it does."""
+    query = parse_qs(urlsplit(url).query)
+    return (query.get("stepname") or query.get("page") or [None])[0]
+
+
 def _apply_url(page: Page) -> str:
     """The wizard's URL: the page's, or where the job page's Apply link leads; the wizard's
     own when no link shows, which `_open` then finds or not."""

{atsfill-0.0.5.dev181 → atsfill-0.0.5.dev184}/src/atsfill/adapters/recruitee.py

@@ -62,6 +62,10 @@ _READ_BACK = """(form, kinds) => {
 }"""
 
 
+def _on_form(url: str) -> bool:
+    return urlsplit(url).path.rstrip("/").endswith("/c/new")
+
+
 class Recruitee:
     name = "recruitee"
     known_gaps: dict[str, Gap] = {}
@@ -84,9 +88,12 @@ class Recruitee:
         match = JOB.match(parts.path)
         return f"https://{parts.hostname}/o/{match[1]}" if match else url
 
+    def open_form(self, page: Page, url: str) -> None:
+        page.goto(url if _on_form(url) else f"{self.posting(url)}/c/new")
+
     def browser_read(self, page: Page) -> list[Field]:
-        if not urlsplit(page.url).path.rstrip("/").endswith("/c/new"):
-            page.goto(f"{self.posting(page.url)}/c/new")
+        if not _on_form(page.url):
+            self.open_form(page, page.url)
         form = page.locator(FORM)
         try:
             form.wait_for(state="attached", timeout=FORM_TIMEOUT_MS)

{atsfill-0.0.5.dev181 → atsfill-0.0.5.dev184}/src/atsfill/adapters/teamtailor.py

@@ -125,6 +125,7 @@ class Teamtailor:
         self._follow_ups: set[str] = set()
         self._uploads: dict[str, str] = {}  # file field id: the answer's file name
         self._videos: dict[str, tuple[str, str]] = {}  # video field id: its video id, its URL
+        self._form: str | None = None  # the form `open_form` loaded
 
     def detect(self, url: str, page: Page | None) -> bool:
         """A `*.teamtailor.com` job, or a job page with Teamtailor's "powered by" link or its
@@ -141,13 +142,21 @@ class Teamtailor:
         assert match, url
         return f"https://{parts.hostname}/jobs/{match[1]}"
 
-    def browser_read(self, page: Page) -> list[Field]:
+    def open_form(self, page: Page, url: str) -> None:
         """Opens the form's own page, never the overlay nor "Apply with LinkedIn"."""
-        parts = urlsplit(page.url)
+        parts = urlsplit(url)
         job = JOB.search(parts.path)
-        assert job, page.url  # detected on a job URL
-        page.goto(f"{parts.scheme}://{parts.netloc}{parts.path[: job.end()]}/applications/new")
-        if urlsplit(page.url).hostname != parts.hostname:
+        assert job, url  # detected on a job URL
+        form = f"{parts.scheme}://{parts.netloc}{parts.path[: job.end()]}/applications/new"
+        page.goto(form)
+        self._form = form
+
+    def browser_read(self, page: Page) -> list[Field]:
+        if self._form is None:  # a page claimed once loaded: a custom domain
+            self.open_form(page, page.url)
+        assert self._form
+        opened, self._form = self._form, None
+        if urlsplit(page.url).hostname != urlsplit(opened).hostname:
             raise Unsupported(f"external_step: the job applies on another site ({page.url})")
         form = page.locator(FORM)
         try:

{atsfill-0.0.5.dev181 → atsfill-0.0.5.dev184}/src/atsfill/adapters/workable.py

@@ -141,6 +141,10 @@ _READ_BACK = """(form, kinds) => {
 }"""
 
 
+def _on_form(url: str) -> bool:
+    return "/apply" in urlsplit(url).path
+
+
 class Workable:
     name = "workable"
     known_gaps = {"entries": Gap.RENDERING}
@@ -174,9 +178,12 @@ class Workable:
         except OSError, ValueError, KeyError, TypeError:
             return None
 
+    def open_form(self, page: Page, url: str) -> None:
+        page.goto(url if _on_form(url) else f"{self.posting(url)}/apply/")
+
     def browser_read(self, page: Page) -> list[Field]:
-        if "/apply" not in urlsplit(page.url).path:
-            page.goto(f"{self.posting(page.url)}/apply/")
+        if not _on_form(page.url):
+            self.open_form(page, page.url)
         form = page.locator(FORM)
         try:
             form.wait_for(timeout=FORM_TIMEOUT_MS)

{atsfill-0.0.5.dev181 → atsfill-0.0.5.dev184}/src/atsfill/core/adapter.py

@@ -280,6 +280,16 @@ class SendsToOpen(Adapter, Protocol):
         ...
 
 
+@runtime_checkable
+class OpensForm(Adapter, Protocol):
+    """An adapter whose form is another page than the posting: the core has it load the form
+    in place of the posting, when it claims the posting by its URL (ADR 0003)."""
+
+    def open_form(self, page: Page, url: str) -> None:
+        """Loads the form of the posting at `url`, so that `browser_read` finds it open."""
+        ...
+
+
 @runtime_checkable
 class ReadsByRequest(Adapter, Protocol):
     """An adapter with a request read: a form's fields listed without a browser (ADR 0011)."""

{atsfill-0.0.5.dev181 → atsfill-0.0.5.dev184}/src/atsfill/core/commands.py

@@ -29,6 +29,7 @@ from atsfill.core.adapter import (
     ListsShownFields,
     NeedsAccount,
     NotTaken,
+    OpensForm,
     ProfileIncomplete,
     ReadsByRequest,
     Refused,
@@ -173,7 +174,7 @@ def _opened(
             session = None
     with open_page(data, session, seconds) as page:
         try:
-            page.goto(url)
+            _load(page, adapter, url)
         except browser.Timeout:
             if adapter is None:
                 raise
@@ -207,6 +208,14 @@ def _opened(
                     browser.save_session(page, session)
 
 
+def _load(page: Page, adapter: object, url: str) -> None:
+    """Loads the posting, or its form where the adapter that claims it opens it (ADR 0003)."""
+    if isinstance(adapter, OpensForm):
+        adapter.open_form(page, url)
+    else:
+        page.goto(url)
+
+
 _again: ContextVar[Callable[[], Result | None] | None] = ContextVar("again", default=None)
 
 
@@ -281,7 +290,7 @@ def _sign_in_again(
             return _login_required(url, adapter, user)
         session = accounts.session_file(mine, adapter.name, scope)
         accounts.sign_in(page, adapter, scope, *credentials, session)
-        page.goto(url)
+        _load(page, adapter, url)
         return None
     email = credentials[0] if credentials else ""
     scoped = url if adapter.per_employer else None