feat: initialize bangwen recruitment aggregation and search engine

- Rust core engine with HTTP ingestion gateway, hybrid search, and white-box resume matching - Python edge collectors supporting declarative TOML and specialized adapters - PostgreSQL schema migrations with pgvector embedding support - Hardened security controls: strict CSP, terminal sanitization, and input boundary validation - Multi-service container orchestration with dedicated non-root execution
2026-10-23 23:44:10 +08:00
commit 1059be1b62
61 changed files with 10489 additions and 0 deletions
@@ -0,0 +1,101 @@
+import json
+import os
+import sys
+import time
+
+import httpx
+from pydantic import BaseModel, Field
+
+
+class AttachmentItem(BaseModel):
+    name: str | None = None
+    url: str
+    mime: str | None = None
+
+
+class PositionItem(BaseModel):
+    post_code: str | None = None
+    dept: str | None = None
+    post_name: str
+    headcount: int | None = None
+    education_min: str | None = None
+    degree_required: bool | None = None
+    majors: list[str] | None = None
+    major_unlimited: bool | None = False
+    age_max: int | None = None
+    bianzhi: str | None = None
+    salary_min: int | None = None
+    salary_max: int | None = None
+    work_city: str | None = None
+    other_req: str | None = None
+    extract_method: str = "edge_python"
+    confidence: float | None = 1.0
+
+
+class AnnouncementItem(BaseModel):
+    url: str
+    url_hash: str | None = None
+    title: str
+    kind: str = "招聘"
+    employer: str | None = None
+    employer_type: str | None = None
+    published_at: str | None = None
+    apply_start: str | None = None
+    apply_end: str | None = None
+    region_hint: str | None = None
+    body_text: str | None = None
+    attachments: list[AttachmentItem] = Field(default_factory=list)
+    positions: list[PositionItem] = Field(default_factory=list)
+    reconcile_positions: bool | None = None
+
+
+class CollectorMeta(BaseModel):
+    name: str
+    runtime: str | None = "python/3.12"
+    auth_status: str = "none"
+    duration_ms: int | None = None
+    error_message: str | None = None
+
+
+class IngestPayload(BaseModel):
+    protocol_version: str = "bangwen/v1"
+    source_id: str
+    batch_id: str
+    collector: CollectorMeta | None = None
+    items: list[AnnouncementItem]
+
+
+def emit_payload(
+    payload: IngestPayload,
+    push_url: str | None = None,
+    token: str | None = None,
+) -> None:
+    data = payload.model_dump(mode="json")
+    if push_url:
+        token = token or os.environ.get("INGEST_TOKEN")
+        headers = {}
+        if token:
+            headers["Authorization"] = f"Bearer {token.strip()}"
+        max_retries = 3
+        last_exc = None
+        for attempt in range(1, max_retries + 1):
+            try:
+                resp = httpx.post(push_url, json=data, headers=headers, timeout=30.0)
+                resp.raise_for_status()
+                sys.stderr.write(
+                    f"Pushed {len(payload.items)} items to {push_url}: {resp.text}\n"
+                )
+                return
+            except Exception as e:
+                last_exc = e
+                sys.stderr.write(
+                    f"Attempt {attempt}/{max_retries} failed to push to {push_url}: {e}\n"
+                )
+                if attempt < max_retries:
+                    time.sleep(2 ** attempt)
+        if last_exc:
+            raise last_exc
+    else:
+        json_output = json.dumps(data, ensure_ascii=False)
+        sys.stdout.write(json_output + "\n")
+        sys.stdout.flush()