fix: separate news from publications and add employee refresh
This commit is contained in:
24
app/admin.py
24
app/admin.py
@@ -17,6 +17,7 @@ from app.services.admin_data import (
|
||||
stats_payload,
|
||||
)
|
||||
from app.services.crawl_control import get_running_run, run_crawl_if_idle
|
||||
from app.services.crawler import refresh_employee
|
||||
from app.version import BACKEND_VERSION, FRONTEND_VERSION
|
||||
|
||||
router = APIRouter(prefix="/admin")
|
||||
@@ -144,10 +145,31 @@ def employee_detail(
|
||||
return _render(
|
||||
request,
|
||||
"employee_detail.html",
|
||||
{"employee": employee, "employee_view": employee_detail_payload(employee), "snapshots": snapshots},
|
||||
{
|
||||
"employee": employee,
|
||||
"employee_view": employee_detail_payload(employee),
|
||||
"snapshots": snapshots,
|
||||
"refresh_status": request.query_params.get("refresh_status"),
|
||||
},
|
||||
)
|
||||
|
||||
|
||||
@router.post("/employees/{employee_id}/refresh")
|
||||
def refresh_employee_detail(
|
||||
employee_id: int,
|
||||
request: Request,
|
||||
db: Session = Depends(get_db),
|
||||
settings: Settings = Depends(get_settings),
|
||||
):
|
||||
require_admin(request, settings)
|
||||
employee = db.get(Employee, employee_id)
|
||||
if not employee:
|
||||
return RedirectResponse("/admin/directory", status_code=303)
|
||||
run = refresh_employee(db, employee, settings)
|
||||
status = "success" if run.status == "completed" else "error"
|
||||
return RedirectResponse(f"/admin/employees/{employee_id}?refresh_status={status}", status_code=303)
|
||||
|
||||
|
||||
@router.get("/runs", response_class=HTMLResponse)
|
||||
def runs(request: Request, db: Session = Depends(get_db), settings: Settings = Depends(get_settings)):
|
||||
require_admin(request, settings)
|
||||
|
||||
@@ -387,7 +387,7 @@ def _infer_section_type(title: str, nodes: list) -> str:
|
||||
lowered = title.lower()
|
||||
if _has_table(nodes):
|
||||
return "table"
|
||||
if "публикац" in lowered:
|
||||
if _is_publications_title(lowered):
|
||||
return "publications"
|
||||
if "учебные курсы" in lowered:
|
||||
return "courses_by_year"
|
||||
@@ -398,6 +398,10 @@ def _infer_section_type(title: str, nodes: list) -> str:
|
||||
return "generic"
|
||||
|
||||
|
||||
def _is_publications_title(lowered_title: str) -> bool:
|
||||
return lowered_title.startswith("публикац")
|
||||
|
||||
|
||||
def _has_table(nodes: list) -> bool:
|
||||
return any(isinstance(node, Tag) and (node.name == "table" or node.find("table")) for node in nodes)
|
||||
|
||||
|
||||
@@ -80,6 +80,48 @@ def run_crawl(db: Session, settings: Settings) -> CrawlRun:
|
||||
return run
|
||||
|
||||
|
||||
def refresh_employee(db: Session, employee: Employee, settings: Settings) -> CrawlRun:
|
||||
run = CrawlRun(source_url=employee.canonical_url, status="running", found_count=1)
|
||||
db.add(run)
|
||||
db.commit()
|
||||
db.refresh(run)
|
||||
|
||||
try:
|
||||
with requests.Session() as session:
|
||||
parsed = parse_person_profile(
|
||||
session,
|
||||
employee.canonical_url,
|
||||
HEADERS,
|
||||
settings.request_timeout,
|
||||
settings.parser_use_playwright,
|
||||
)
|
||||
if not parsed:
|
||||
raise ValueError("Профиль не удалось распарсить.")
|
||||
if _parsed_profile_key(parsed) != employee.profile_key:
|
||||
raise ValueError("Распарсенный профиль не совпадает с обновляемым сотрудником.")
|
||||
|
||||
_upsert_employee(db, run, parsed)
|
||||
run.parsed_count = 1
|
||||
run.status = "completed"
|
||||
except Exception as exc:
|
||||
run.status = "failed"
|
||||
run.error_count = 1
|
||||
run.message = str(exc)
|
||||
db.add(
|
||||
CrawlError(
|
||||
crawl_run_id=run.id,
|
||||
profile_url=employee.canonical_url,
|
||||
error_type=type(exc).__name__,
|
||||
message=str(exc),
|
||||
)
|
||||
)
|
||||
finally:
|
||||
run.finished_at = datetime.now(timezone.utc)
|
||||
db.commit()
|
||||
db.refresh(run)
|
||||
return run
|
||||
|
||||
|
||||
def _ensure_source(db: Session, source_url: str) -> ParserSource:
|
||||
source = db.scalar(select(ParserSource).where(ParserSource.source_url == source_url))
|
||||
if source:
|
||||
@@ -91,10 +133,14 @@ def _ensure_source(db: Session, source_url: str) -> ParserSource:
|
||||
return source
|
||||
|
||||
|
||||
def _parsed_profile_key(parsed: dict) -> str:
|
||||
return f"{parsed.get('profile_type')}:{parsed.get('profile_id')}"
|
||||
|
||||
|
||||
def _upsert_employee(db: Session, run: CrawlRun, parsed: dict) -> Employee:
|
||||
html = parsed.pop("_html", None)
|
||||
checksum = _checksum(parsed)
|
||||
key = f"{parsed.get('profile_type')}:{parsed.get('profile_id')}"
|
||||
key = _parsed_profile_key(parsed)
|
||||
employee = db.scalar(select(Employee).where(Employee.profile_key == key))
|
||||
now = datetime.now(timezone.utc)
|
||||
if not employee:
|
||||
|
||||
@@ -171,6 +171,10 @@
|
||||
background: transparent;
|
||||
}
|
||||
|
||||
.button--compact {
|
||||
padding: 8px 12px;
|
||||
}
|
||||
|
||||
.code {
|
||||
overflow-x: auto;
|
||||
padding: 14px;
|
||||
@@ -201,11 +205,34 @@
|
||||
gap: 10px;
|
||||
}
|
||||
|
||||
.employee-card__actions {
|
||||
display: grid;
|
||||
justify-items: end;
|
||||
gap: 10px;
|
||||
}
|
||||
|
||||
.employee-card__title {
|
||||
margin: 0;
|
||||
font-size: 24px;
|
||||
}
|
||||
|
||||
.employee-card__notice {
|
||||
margin: 0;
|
||||
padding: 12px 14px;
|
||||
border-radius: 8px;
|
||||
font-weight: 700;
|
||||
}
|
||||
|
||||
.employee-card__notice--success {
|
||||
color: #065f46;
|
||||
background: #d1fae5;
|
||||
}
|
||||
|
||||
.employee-card__notice--error {
|
||||
color: #991b1b;
|
||||
background: #fee2e2;
|
||||
}
|
||||
|
||||
.employee-card__section {
|
||||
padding: 20px;
|
||||
background: #ffffff;
|
||||
|
||||
@@ -7,8 +7,18 @@
|
||||
<h2 class="employee-card__title">{{ employee_view.full_name or employee.profile_key }}</h2>
|
||||
<span class="badge {% if employee_view.status == "dismissed" %}badge--dismissed{% endif %}">{{ employee_view.status_display }}</span>
|
||||
</div>
|
||||
<a class="admin__link" href="{{ employee_view.canonical_url }}">{{ employee_view.canonical_url }}</a>
|
||||
<div class="employee-card__actions">
|
||||
<form method="post" action="/admin/employees/{{ employee.id }}/refresh">
|
||||
<button class="button button--compact" type="submit">Обновить данные</button>
|
||||
</form>
|
||||
<a class="admin__link" href="{{ employee_view.canonical_url }}">{{ employee_view.canonical_url }}</a>
|
||||
</div>
|
||||
</div>
|
||||
{% if refresh_status == "success" %}
|
||||
<p class="employee-card__notice employee-card__notice--success">Данные сотрудника обновлены.</p>
|
||||
{% elif refresh_status == "error" %}
|
||||
<p class="employee-card__notice employee-card__notice--error">Не удалось обновить данные сотрудника.</p>
|
||||
{% endif %}
|
||||
|
||||
<section class="employee-card__section">
|
||||
<h3 class="employee-section__title">Основная информация</h3>
|
||||
|
||||
@@ -1,3 +1,3 @@
|
||||
APP_VERSION = "0.4.6"
|
||||
FRONTEND_VERSION = "0.4.6"
|
||||
BACKEND_VERSION = "0.4.6"
|
||||
APP_VERSION = "0.4.7"
|
||||
FRONTEND_VERSION = "0.4.7"
|
||||
BACKEND_VERSION = "0.4.7"
|
||||
|
||||
Reference in New Issue
Block a user