Read a search role's own page for its text in enrich and on the Fetch button when its reader has detail.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
This commit is contained in:
1 parent
07e3bb4756
commit
b148647ac3
5 files changed
+346
-40
No files matched your search
+51
-31
@@ -2,18 +2,22 @@
|
||||
|
||||
Some hosts list a role without its text and state it on the role's own page
|
||||
only (`text="detail"`: Workday, SuccessFactors, Oracle, Phenom and the rest
|
||||
with a `detail`). The pull reads listings only; this stage reads those pages
|
||||
after scoring, so the calls go where the ranking is:
|
||||
with a `detail`, and a search provider such as Jobbank). The pull reads
|
||||
listings only; this stage reads those pages after scoring, so the calls go
|
||||
where the ranking is. A saved search whose reader has `detail` is treated as
|
||||
a board under its target `provider:name`: the same bar, budgets and state
|
||||
keys, and its failing record is `search:failing:<target>`.
|
||||
|
||||
1. `enrich_waiting` hands over the open roles without text that an enabled
|
||||
agent with a prompt ranks ungated among its best `enrich.top` and at
|
||||
`enrich.min_score` or above, best first by the highest such score. A role
|
||||
below that bar for every agent gets no call.
|
||||
2. They are taken in that order, those whose read failed last, on boards whose reader has `detail`, at most
|
||||
`enrich.per_board` from one board and `enrich.per_run` in all, leaving
|
||||
out the roles already read, unless their read failed or gave no text
|
||||
`enrich.retry_days` ago (`plugins.ats.detail_due`).
|
||||
3. One batch per board, on the `http` pool, reads them
|
||||
2. They are taken in that order, those whose read failed last, on boards and
|
||||
searches whose reader has `detail`, at most `enrich.per_board` from one
|
||||
board and `enrich.per_run` in all, leaving out the roles already read,
|
||||
unless their read failed or gave no text `enrich.retry_days` ago
|
||||
(`plugins.ats.detail_due`).
|
||||
3. One batch per board or search, on the `http` pool, reads them
|
||||
(`plugins.ats.read_board_details`, which keeps `ats:detailed:<target>`,
|
||||
`ats:detailfailed:<target>` and `ats:detailempty:<target>`) and stores
|
||||
what each page filled. `Store.upsert_role` drops the vector of a role
|
||||
@@ -37,6 +41,7 @@ import httpx
|
||||
|
||||
from app.config import Config
|
||||
from app.jobs import stages
|
||||
from app.models import Subscription
|
||||
from app.storage import Store
|
||||
from pipeline import NULL, Batch, RunContext, Tally, Task, job
|
||||
|
||||
@@ -53,7 +58,14 @@ def _ats() -> Any:
|
||||
return ats
|
||||
|
||||
|
||||
def _row(found: dict[str, Any], target: str) -> dict[str, Any]:
|
||||
def _native(found: dict[str, Any], sub: Subscription) -> str:
|
||||
"""The host's own id of a stored role: a board's roles are stored as
|
||||
`host:slug:id`, a search's as `provider:id`."""
|
||||
prefix = sub.target if sub.source == "ats" else sub.target.partition(":")[0]
|
||||
return str(found["source_id"]).removeprefix(f"{prefix}:")
|
||||
|
||||
|
||||
def _row(found: dict[str, Any], sub: Subscription) -> dict[str, Any]:
|
||||
"""A stored role back as the row its reader listed: the stored `raw`
|
||||
with the host's own id, the title, the link and the date."""
|
||||
try:
|
||||
@@ -65,29 +77,36 @@ def _row(found: dict[str, Any], target: str) -> dict[str, Any]:
|
||||
posted: datetime | None = datetime.fromisoformat(str(found.get("published_at") or ""))
|
||||
except ValueError:
|
||||
posted = None
|
||||
return raw | {"source_id": str(found["source_id"]).removeprefix(f"{target}:"),
|
||||
return raw | {"source_id": _native(found, sub),
|
||||
"title": found["title"], "url": found["url"], "description": "",
|
||||
"posted": posted}
|
||||
|
||||
|
||||
def _readable(store: Store) -> dict[int, Subscription]:
|
||||
"""The fetchable boards and searches whose reader has `detail`, by
|
||||
subscription id."""
|
||||
ats = _ats()
|
||||
return {sub.id: sub for source in ("ats", "search")
|
||||
for sub in store.subscriptions(source=source, fetchable=True)
|
||||
if hasattr(ats.reader_of(sub.target), "detail")}
|
||||
|
||||
|
||||
def plan(config: Config, store: Store) -> list[tuple[str, list[str]]]:
|
||||
"""`(target, uids)` per board, best board first, within the budgets: the
|
||||
roles `enrich_waiting` ranks, on fetchable boards whose reader has
|
||||
`detail` and that are not failing, and due for a read."""
|
||||
"""`(target, uids)` per board or search, best first, within the budgets:
|
||||
the roles `enrich_waiting` ranks, on fetchable boards and searches whose
|
||||
reader has `detail` and that are not failing, and due for a read."""
|
||||
settings = config.enrich
|
||||
ats = _ats()
|
||||
boards = {sub.id: sub.target for sub in store.subscriptions(source="ats", fetchable=True)}
|
||||
subs = _readable(store)
|
||||
by_board: dict[str, list[dict[str, Any]]] = defaultdict(list)
|
||||
for found in store.enrich_waiting(settings.top, settings.min_score):
|
||||
target = boards.get(found["subscription_id"])
|
||||
if target is None:
|
||||
sub = subs.get(found["subscription_id"])
|
||||
if sub is None or f"{sub.source}:failing:{sub.target}" in store.state:
|
||||
continue
|
||||
module = ats.readers.HOSTS.get(target.partition(":")[0])
|
||||
if module is None or not hasattr(module, "detail") \
|
||||
or f"ats:failing:{target}" in store.state:
|
||||
continue
|
||||
by_board[target].append(found)
|
||||
due = {target: set(ats.detail_due(store.state, target, [f["source_id"] for f in found],
|
||||
# The id the board's per-role state keeps: `host:slug:id`.
|
||||
found["compound"] = f"{sub.target}:{_native(found, sub)}"
|
||||
by_board[sub.target].append(found)
|
||||
due = {target: set(ats.detail_due(store.state, target, [f["compound"] for f in found],
|
||||
retry_days=settings.retry_days))
|
||||
for target, found in by_board.items()}
|
||||
# Best first; roles whose read failed after every other, so a page that
|
||||
@@ -95,13 +114,13 @@ def plan(config: Config, store: Store) -> list[tuple[str, list[str]]]:
|
||||
failed = {target: ats._runner.id_set(store.state, ats.detail_keys(target)[1])
|
||||
for target in by_board}
|
||||
ranked = sorted((f for found in by_board.values() for f in found),
|
||||
key=lambda f: (f["source_id"] in failed[boards[f["subscription_id"]]],
|
||||
key=lambda f: (f["compound"] in failed[subs[f["subscription_id"]].target],
|
||||
-float(f["score"] or 0), f["uid"]))
|
||||
picked: dict[str, list[str]] = {}
|
||||
total = 0
|
||||
for found in ranked:
|
||||
target = boards[found["subscription_id"]]
|
||||
if found["source_id"] not in due[target]:
|
||||
target = subs[found["subscription_id"]].target
|
||||
if found["compound"] not in due[target]:
|
||||
continue
|
||||
if total >= settings.per_run:
|
||||
break
|
||||
@@ -114,7 +133,8 @@ def plan(config: Config, store: Store) -> list[tuple[str, list[str]]]:
|
||||
|
||||
|
||||
def batches(config: Config, store: Store) -> list[Batch]:
|
||||
"""The finder: one batch per board, nothing while the stage is off."""
|
||||
"""The finder: one batch per board or search, nothing while the stage is
|
||||
off."""
|
||||
if not stages.stage_on(config, store, "enrich"):
|
||||
return []
|
||||
return [Batch({"target": target, "uids": uids},
|
||||
@@ -125,16 +145,16 @@ def batches(config: Config, store: Store) -> list[Batch]:
|
||||
@job(KIND, "Reading roles on their own pages", resource="http")
|
||||
def run_batch(config: Config, store: Store, *, target: str, uids: list[str],
|
||||
context: RunContext = NULL) -> dict[str, Any]:
|
||||
"""Read one board's roles in `uids` on their own pages, in that order,
|
||||
and store what each page filled. `filled` is how many roles changed,
|
||||
`texts` how many of them now have text."""
|
||||
"""Read one board's or search's roles in `uids` on their own pages, in
|
||||
that order, and store what each page filled. `filled` is how many roles
|
||||
changed, `texts` how many of them now have text."""
|
||||
ats = _ats()
|
||||
subs = [s for s in store.subscriptions(source="ats", fetchable=True) if s.target == target]
|
||||
subs = [s for s in _readable(store).values() if s.target == target]
|
||||
if not subs or not uids:
|
||||
return {"target": target, "filled": 0, "texts": 0}
|
||||
sub = subs[0]
|
||||
found = {f["uid"]: f for f in store.enrich_waiting(0, 0.0, uids)}
|
||||
rows = [_row(found[uid], target) for uid in uids if uid in found]
|
||||
rows = [_row(found[uid], sub) for uid in uids if uid in found]
|
||||
context.check()
|
||||
with httpx.Client() as client:
|
||||
filled = ats.read_board_details(store.state, target, rows, client=client,
|
||||
@@ -143,7 +163,7 @@ def run_batch(config: Config, store: Store, *, target: str, uids: list[str],
|
||||
texts = 0
|
||||
for row in filled:
|
||||
item = ats.item_of(row, target, label)
|
||||
item.source = "ats"
|
||||
item.source = sub.source
|
||||
store.upsert_role(item, sub)
|
||||
texts += bool(row.get("description"))
|
||||
log.info("enrich: %s: %d read, %d filled, %d with text", target, len(rows), len(filled),
|
||||
|
||||
+51
-2
@@ -5,6 +5,11 @@ link, keep the text as `article_text`. It is on demand, never in the pull,
|
||||
because a fetch of a Jobindex page counts as a click in Jobindex's statistics
|
||||
and one click for a posting somebody opened is fair.
|
||||
|
||||
A role of a search whose reader has `detail` (Jobbank) is read through that
|
||||
`detail`, which reads the ad's own page and states its text, rather than
|
||||
scraping the link; a failure of that read is an `Unreadable` with the
|
||||
category its HTTP status or exception names.
|
||||
|
||||
A Jobindex role links to Jobindex's own `vis-job` page, which carries a short
|
||||
ad and a "Se jobbet" button to the posting itself: the employer's site, The
|
||||
Hub or Jobindex's `jobannonce`. That button's link is found, kept as
|
||||
@@ -19,9 +24,14 @@ message the route answers with.
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
from datetime import datetime
|
||||
from typing import Any
|
||||
|
||||
import httpx
|
||||
|
||||
from app.models import Role
|
||||
from app.storage import Store
|
||||
from desk.tools import scrape
|
||||
from desk.tools import failures, scrape
|
||||
from desk.tools.checks import Page
|
||||
from plugins.readers import jobindex
|
||||
|
||||
@@ -113,6 +123,43 @@ def _jobindex_text(store: Store, role: Role) -> str:
|
||||
raise Unreadable(failure or own.failure)
|
||||
|
||||
|
||||
def _detail_target(role: Role) -> str:
|
||||
"""The search target of a role whose reader has `detail`, else ''."""
|
||||
from plugins import ats
|
||||
|
||||
target = str(role.raw.get("search") or "")
|
||||
if role.source == "search" and hasattr(ats.reader_of(target), "detail"):
|
||||
return target
|
||||
return ""
|
||||
|
||||
|
||||
def _detail_text(role: Role, target: str) -> str:
|
||||
"""The text of a search role, read by its reader's `detail` on the ad's
|
||||
own page as the pull's enrich stage reads it. Raises `Unreadable` when
|
||||
the read fails or the page states no text."""
|
||||
from plugins import ats
|
||||
|
||||
provider = target.partition(":")[0]
|
||||
try:
|
||||
posted: datetime | None = datetime.fromisoformat(role.published_at)
|
||||
except ValueError:
|
||||
posted = None
|
||||
row: dict[str, Any] = role.raw | {
|
||||
"source_id": role.source_id.removeprefix(f"{provider}:"), "title": role.title,
|
||||
"url": role.original_url, "description": "", "posted": posted}
|
||||
try:
|
||||
with httpx.Client() as client:
|
||||
found = ats.read_detail(client, target, row)
|
||||
except httpx.HTTPStatusError as exc:
|
||||
raise Unreadable(failures.of_status(exc.response.status_code)) from exc
|
||||
except Exception as exc:
|
||||
category = failures.of_exception(exc)
|
||||
raise Unreadable("" if category == failures.UNKNOWN else category) from exc
|
||||
if not (text := str(found.get("description") or "").strip()):
|
||||
raise Unreadable("empty")
|
||||
return text
|
||||
|
||||
|
||||
def fetch_posting(store: Store, uid: str) -> Role:
|
||||
"""Fetch the posting page once and keep it as plain text, for a role
|
||||
whose description is a teaser. Synchronous: one request the person asked
|
||||
@@ -124,7 +171,9 @@ def fetch_posting(store: Store, uid: str) -> Role:
|
||||
raise AlreadyFetched("The posting text is already fetched.")
|
||||
if not role.original_url:
|
||||
raise NoLink("This role has no link to fetch the posting from.")
|
||||
if _is_jobindex(role):
|
||||
if target := _detail_target(role):
|
||||
text = _detail_text(role, target)
|
||||
elif _is_jobindex(role):
|
||||
text = _jobindex_text(store, role)
|
||||
else:
|
||||
got = scrape.scrape(role.original_url)
|
||||
|
||||
Reference in new issue
Block a user