flru-parser 0.3.0__py3-none-any.whl

flru/__init__.py

@@ -0,0 +1,138 @@
+from .batch import BatchError, BatchResult
+from .client import FLClient
+from .config import (
+    CircuitBreakerConfig,
+    ClientConfig,
+    ProxyConfig,
+    RateLimitConfig,
+    RetryConfig,
+    TimeoutConfig,
+)
+from .easy import (
+    Client,
+    Pages,
+    fetch_freelancers,
+    fetch_project,
+    fetch_projects,
+    fetch_user,
+)
+from .exceptions import (
+    AuthenticationRequired,
+    BlockedError,
+    CircuitOpenError,
+    ConfigurationError,
+    EmptyPageError,
+    FLRUError,
+    HTTPStatusError,
+    ParseError,
+    RateLimitedError,
+    RobotsDeniedError,
+    SecurityError,
+    SelectorDriftError,
+    TransportError,
+    UnexpectedPageError,
+)
+from .filters import ProjectFilters, ProjectType
+from .models import (
+    Attachment,
+    Category,
+    CrawlCheckpoint,
+    FreelancerPage,
+    FreelancerSummary,
+    Image,
+    Link,
+    Money,
+    PageData,
+    ParseDiagnostics,
+    PortfolioItem,
+    ProjectDetail,
+    ProjectKind,
+    ProjectPage,
+    ProjectRecord,
+    ProjectStatus,
+    ProjectSummary,
+    RequestMetrics,
+    Review,
+    SourceInfo,
+    UserProfile,
+    UserSummary,
+)
+from .observability import RequestEvent, StructuredLogHandler
+from .state import (
+    CrawlStateStore,
+    MemoryStateStore,
+    PostgresStateStore,
+    RedisStateStore,
+    SQLiteStateStore,
+    project_content_hash,
+)
+
+__version__ = "0.3.0"
+
+AsyncClient = Client
+
+__all__ = [
+    "__version__",
+    "AsyncClient",
+    "Attachment",
+    "AuthenticationRequired",
+    "BatchError",
+    "BatchResult",
+    "BlockedError",
+    "Category",
+    "CircuitBreakerConfig",
+    "CircuitOpenError",
+    "Client",
+    "ClientConfig",
+    "ConfigurationError",
+    "CrawlCheckpoint",
+    "CrawlStateStore",
+    "EmptyPageError",
+    "FLClient",
+    "FLRUError",
+    "FreelancerPage",
+    "FreelancerSummary",
+    "HTTPStatusError",
+    "Image",
+    "Link",
+    "MemoryStateStore",
+    "Money",
+    "PageData",
+    "Pages",
+    "ParseDiagnostics",
+    "ParseError",
+    "PortfolioItem",
+    "PostgresStateStore",
+    "ProjectDetail",
+    "ProjectFilters",
+    "ProjectKind",
+    "ProjectPage",
+    "ProjectRecord",
+    "ProjectStatus",
+    "ProjectSummary",
+    "ProjectType",
+    "ProxyConfig",
+    "RateLimitConfig",
+    "RateLimitedError",
+    "RedisStateStore",
+    "RequestEvent",
+    "RequestMetrics",
+    "RetryConfig",
+    "Review",
+    "RobotsDeniedError",
+    "SQLiteStateStore",
+    "SecurityError",
+    "SelectorDriftError",
+    "SourceInfo",
+    "StructuredLogHandler",
+    "TimeoutConfig",
+    "TransportError",
+    "UnexpectedPageError",
+    "UserProfile",
+    "UserSummary",
+    "project_content_hash",
+    "fetch_freelancers",
+    "fetch_project",
+    "fetch_projects",
+    "fetch_user",
+]

flru/batch.py

@@ -0,0 +1,28 @@
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from typing import Generic, TypeVar
+
+T = TypeVar("T")
+K = TypeVar("K")
+
+
+@dataclass(slots=True, frozen=True)
+class BatchError(Generic[K]):
+    key: K
+    error: Exception
+
+
+@dataclass(slots=True)
+class BatchResult(Generic[K, T]):
+    successful: list[T] = field(default_factory=list)
+    failed: list[BatchError[K]] = field(default_factory=list)
+
+    @property
+    def ok(self) -> bool:
+        return not self.failed
+
+    def raise_for_errors(self) -> None:
+        if self.failed:
+            first = self.failed[0]
+            raise first.error

flru/canary.py

@@ -0,0 +1,33 @@
+from __future__ import annotations
+
+import argparse
+import asyncio
+import json
+
+from .client import FLClient
+
+
+async def run_canary() -> dict[str, object]:
+    async with FLClient() as client:
+        page = await client.get_projects_page(1)
+        first = page.items[0]
+        return {
+            "ok": True,
+            "items": len(page.items),
+            "first_id": first.id,
+            "first_title": first.title,
+            "fingerprint": page.diagnostics.page_fingerprint,
+            "confidence": page.diagnostics.confidence,
+        }
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(description="Run a read-only FL.ru parser canary")
+    parser.add_argument("--pretty", action="store_true")
+    args = parser.parse_args()
+    result = asyncio.run(run_canary())
+    print(json.dumps(result, ensure_ascii=False, indent=2 if args.pretty else None))
+
+
+if __name__ == "__main__":
+    main()