fix: staticmethod for repos collecting

Author: jbristow

Makefile

@@ -20,8 +20,8 @@ fmt:
 	poetry run isort nodestream_github tests
 	poetry run black nodestream_github tests
 
-.PHONY: fmt
-lint:
+.PHONY: lint
+lint: fmt
 	poetry run ruff check nodestream_github tests --fix
 
 .PHONY: test

nodestream_github/client/githubclient.py

@@ -23,7 +23,7 @@
 import nodestream_github.types as types
 from nodestream_github.logging import get_plugin_logger
 
-DEFAULT_REQUEST_RATE_LIMIT_PER_MINUTE = int(13000 / 60 )
+DEFAULT_REQUEST_RATE_LIMIT_PER_MINUTE = int(13000 / 60)
 DEFAULT_MAX_RETRIES = 20
 DEFAULT_PAGE_SIZE = 100
 

nodestream_github/logging/__init__.py

@@ -1,7 +1,7 @@
 import logging
 
 
-def get_plugin_logger(name:str)->logging.Logger:
+def get_plugin_logger(name: str) -> logging.Logger:
     logger = logging.getLogger(name)
     logger.addHandler(logging.NullHandler())
-    return logger
+    return logger

nodestream_github/repos.py

@@ -6,7 +6,7 @@
 """
 
 from collections.abc import AsyncGenerator
-from dataclasses import InitVar, dataclass, field
+from dataclasses import dataclass
 
 from nodestream.pipeline import Extractor
 
@@ -30,22 +30,12 @@ def _dict_val_to_bool(d: dict[str, any], key: str) -> bool:
 
 @dataclass
 class CollectWhichRepos:
-    org_all: InitVar[bool] = field(default=False)
-    user_all: InitVar[bool] = field(default=False)
     all_public: bool = False
     org_public: bool = False
     org_private: bool = False
     user_public: bool = False
     user_private: bool = False
 
-    def __post_init__(self, org_all: bool, user_all: bool):
-        if org_all:
-            self.org_public = True
-            self.org_private = True
-        if user_all:
-            self.user_public = True
-            self.user_private = True
-
     @property
     def org_any(self) -> bool:
         return self.org_public or self.org_private
@@ -54,6 +44,19 @@ def org_any(self) -> bool:
     def user_any(self) -> bool:
         return self.user_public or self.user_private
 
+    @staticmethod
+    def from_dict(raw_dict: dict[str, any]) -> "CollectWhichRepos":
+        org_all = _dict_val_to_bool(raw_dict, "org_all")
+        user_all = _dict_val_to_bool(raw_dict, "user_all")
+
+        return CollectWhichRepos(
+            all_public=_dict_val_to_bool(raw_dict, "all_public"),
+            org_public=org_all or _dict_val_to_bool(raw_dict, "org_public"),
+            org_private=org_all or _dict_val_to_bool(raw_dict, "org_private"),
+            user_public=user_all or _dict_val_to_bool(raw_dict, "user_public"),
+            user_private=user_all or _dict_val_to_bool(raw_dict, "user_private"),
+        )
+
 
 class GithubReposExtractor(Extractor):
     def __init__(
@@ -64,7 +67,7 @@ def __init__(
         if isinstance(collecting, CollectWhichRepos):
             self.collecting = collecting
         elif isinstance(collecting, dict):
-            self.collecting = CollectWhichRepos(**collecting)
+            self.collecting = CollectWhichRepos.from_dict(collecting)
         else:
             self.collecting = CollectWhichRepos()
         self.client = GithubRestApiClient(**kwargs)

nodestream_github/teams.py

@@ -28,7 +28,11 @@ async def extract_records(self) -> AsyncGenerator[TeamRecord]:
             async for team in self.client.fetch_teams_for_org(login):
                 team_record = await self._fetch_team(login, team)
                 if team_record:
-                    logger.debug("yielded GithubTeam{org=%s,slug=%s}", team_record["organization"]["login"], team_record["slug"], )
+                    logger.debug(
+                        "yielded GithubTeam{org=%s,slug=%s}",
+                        team_record["organization"]["login"],
+                        team_record["slug"],
+                    )
                     yield team_record
 
     async def _fetch_members(self, team: GithubTeam) -> AsyncGenerator[SimplifiedUser]: