openhound-github 0.6.4__tar.gz → 0.6.6__tar.gz

{openhound_github-0.6.4 → openhound_github-0.6.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: openhound-github
-Version: 0.6.4
+Version: 0.6.6
 Summary: Github collector for OpenHound
 Author-email: "jdreijer@specterops.io" <jdreijer@specterops.io>, Jared Atkinson <jatkinson@specterops.io>
 License-File: LICENSE.md

{openhound_github-0.6.4 → openhound_github-0.6.6}/src/openhound_github/helpers.py

@@ -1,11 +1,13 @@
 import logging
 import time
-from typing import Optional
+from collections.abc import Iterator
+from typing import Any, Optional
 from urllib.parse import urlparse
 
 from dlt.common import jsonpath
 from dlt.sources.helpers import requests
 from dlt.sources.helpers.rest_client.auth import AuthConfigBase
+from dlt.sources.helpers.rest_client.client import RESTClient
 from dlt.sources.helpers.rest_client.paginators import (
     JSONResponseCursorPaginator,
 )
@@ -20,6 +22,22 @@ class GraphQLPaginationError(RuntimeError):
     pass
 
 
+class AdaptiveGraphQLPageError(RuntimeError):
+    """A terminal GraphQL page failure with adaptive pagination context."""
+
+    def __init__(
+        self,
+        *,
+        cursor: str | None,
+        page_size: int,
+        error: BaseException,
+    ) -> None:
+        super().__init__(str(error))
+        self.cursor = cursor
+        self.page_size = page_size
+        self.error = error
+
+
 def scim_skip_reason(exception: BaseException) -> str | None:
     """Return a user-facing reason for expected SCIM API unavailability."""
     if not isinstance(exception, requests.HTTPError) or exception.response is None:
@@ -56,6 +74,10 @@ class GraphQLCursorPaginator(JSONResponseCursorPaginator):
         self.has_next_field = has_next_field
         self.allow_missing_page_info = allow_missing_page_info
 
+    @property
+    def next_cursor(self) -> str | None:
+        return self._next_reference
+
     def init_request(self, request: "Request") -> None:
         self._next_reference = None
         self._has_next_page = True
@@ -149,6 +171,156 @@ class GraphQLCursorPaginator(JSONResponseCursorPaginator):
         variables[self.cursor_variable] = self._next_reference
 
 
+def _graphql_gateway_status(exception: BaseException) -> int | None:
+    if not isinstance(exception, requests.HTTPError) or exception.response is None:
+        return None
+
+    status_code = exception.response.status_code
+    if status_code in (502, 504):
+        return status_code
+    return None
+
+
+def adaptive_graphql_paginate(
+    client: RESTClient,
+    *,
+    query: str,
+    variables: dict[str, Any],
+    page_info_path: str,
+    cursor_variable: str = "after",
+    page_size_variable: str = "count",
+    page_sizes: tuple[int, ...] = (100, 50, 25),
+    allow_missing_page_info: bool = False,
+    resource_name: str = "graphql",
+    scope_name: str | None = None,
+    scope_value: str | None = None,
+    log: logging.Logger | None = None,
+) -> Iterator[dict[str, Any]]:
+    """Yield GraphQL pages while adapting connection size after gateway failures.
+
+    The underlying HTTP session still owns ordinary request retries. This helper
+    handles the remaining terminal 502/504 case by retrying the same cursor with
+    progressively smaller connection sizes.
+    """
+    if not page_sizes or any(page_size <= 0 for page_size in page_sizes):
+        raise ValueError("page_sizes must contain positive integers")
+    if tuple(sorted(set(page_sizes), reverse=True)) != page_sizes:
+        raise ValueError("page_sizes must be unique and in descending order")
+
+    logger_instance = log or logger
+    base_page_size = page_sizes[0]
+    cursor = variables.get(cursor_variable)
+    if cursor is not None and not isinstance(cursor, str):
+        raise ValueError(f"{cursor_variable} must be a string or None")
+
+    stable_page_size = base_page_size
+    probe_base_page_size = False
+
+    while True:
+        if stable_page_size == base_page_size:
+            candidate_sizes = page_sizes
+            probing_base_page_size = False
+        elif probe_base_page_size:
+            candidate_sizes = (base_page_size,) + tuple(
+                page_size
+                for page_size in page_sizes
+                if page_size <= stable_page_size
+            )
+            probing_base_page_size = True
+        else:
+            candidate_sizes = tuple(
+                page_size
+                for page_size in page_sizes
+                if page_size <= stable_page_size
+            )
+            probing_base_page_size = False
+
+        page_data: dict[str, Any] | None = None
+        paginator: GraphQLCursorPaginator | None = None
+        successful_page_size: int | None = None
+
+        for index, page_size in enumerate(candidate_sizes):
+            request_variables = {
+                **variables,
+                cursor_variable: cursor,
+                page_size_variable: page_size,
+            }
+            try:
+                response = client.post(
+                    "/graphql",
+                    json={"query": query, "variables": request_variables},
+                )
+                response.raise_for_status()
+                response_json = response.json()
+                paginator = GraphQLCursorPaginator(
+                    page_info_path=page_info_path,
+                    cursor_variable=cursor_variable,
+                    cursor_field="endCursor",
+                    has_next_field="hasNextPage",
+                    allow_missing_page_info=allow_missing_page_info,
+                )
+                paginator.update_state(response)
+                page_data = response_json.get("data")
+                if not isinstance(page_data, dict):
+                    raise GraphQLPaginationError(
+                        "GraphQL response data must be an object"
+                    )
+                successful_page_size = page_size
+                break
+            except Exception as error:
+                status_code = _graphql_gateway_status(error)
+                next_page_size = (
+                    candidate_sizes[index + 1]
+                    if status_code is not None and index + 1 < len(candidate_sizes)
+                    else None
+                )
+                if next_page_size is not None:
+                    scope = (
+                        f" {scope_name} '{scope_value}'"
+                        if scope_name and scope_value
+                        else ""
+                    )
+                    logger_instance.warning(
+                        "Adaptive GraphQL pagination for resource '%s'%s retrying "
+                        "cursor %r with page size %d after HTTP %d at page size %d",
+                        resource_name,
+                        scope,
+                        cursor,
+                        next_page_size,
+                        status_code,
+                        page_size,
+                    )
+                    continue
+                raise AdaptiveGraphQLPageError(
+                    cursor=cursor,
+                    page_size=page_size,
+                    error=error,
+                ) from error
+
+        if page_data is None or paginator is None or successful_page_size is None:
+            raise GraphQLPaginationError("GraphQL page did not produce data")
+
+        yield page_data
+
+        if successful_page_size == base_page_size:
+            stable_page_size = base_page_size
+            probe_base_page_size = False
+        elif stable_page_size == base_page_size:
+            stable_page_size = successful_page_size
+            probe_base_page_size = True
+        elif probing_base_page_size:
+            stable_page_size = successful_page_size
+            probe_base_page_size = False
+        elif successful_page_size < stable_page_size:
+            stable_page_size = successful_page_size
+            probe_base_page_size = False
+
+        if not paginator.has_next_page:
+            break
+
+        cursor = paginator.next_cursor
+
+
 def _response_message(response: requests.Response) -> str:
     try:
         response_data = response.json()

{openhound_github-0.6.4 → openhound_github-0.6.6}/src/openhound_github/resources/organization.py

@@ -22,7 +22,12 @@ from openhound_github.graphql import (
     TEAM_MEMBERS_OVERFLOW_QUERY,
     TEAMS_QUERY,
 )
-from openhound_github.helpers import GraphQLCursorPaginator, scim_skip_reason
+from openhound_github.helpers import (
+    AdaptiveGraphQLPageError,
+    GraphQLCursorPaginator,
+    adaptive_graphql_paginate,
+    scim_skip_reason,
+)
 from openhound_github.main import app
 from openhound_github.models import (
     ActionPermission,
@@ -101,6 +106,7 @@ class SourceContext:
     github_deployment_id: str = DEFAULT_GITHUB_DEPLOYMENT_ID
     github_web_origin: str = DEFAULT_GITHUB_WEB_ORIGIN
     cache_lock: Lock = field(default_factory=Lock)
+    organizations_cache: dict[str, dict[str, Any]] = field(default_factory=dict)
     app_cache: dict[str, dict[str, Any]] = field(default_factory=dict)
     actions_permissions_cache: dict[str, dict[str, Any]] = field(default_factory=dict)
     runner_permissions_cache: dict[str, dict[str, Any]] = field(default_factory=dict)
@@ -252,7 +258,13 @@ def organizations(ctx: SourceContext):
         org_name = org.org_name
         client = org.client
         try:
-            org_data = client.get(f"/orgs/{org_name}").json()
+            org_data = _cached_org_response(
+                ctx,
+                ctx.organizations_cache,
+                client,
+                org_name,
+                f"/orgs/{org_name}",
+            )
 
             actions = _actions_permissions(ctx, client, org_name)
             self_hosted_runners = _runner_permissions(ctx, client, org_name)
@@ -919,25 +931,17 @@ def repositories_graphql(ctx: SourceContext):
         repository_cursor: str | None = None
         emitted_repositories = 0
         try:
-            paginator = GraphQLCursorPaginator(
+            for page_data in adaptive_graphql_paginate(
+                client,
+                query=REPO_REFS_QUERY,
+                variables={"login": org_name, "after": None},
                 page_info_path="data.organization.repositories.pageInfo",
-                cursor_variable="after",
-                cursor_field="endCursor",
-                has_next_field="hasNextPage",
-            )
-            data = {
-                "query": REPO_REFS_QUERY,
-                "variables": {"login": org_name, "count": 100, "after": None},
-            }
-
-            for page_data in client.paginate(
-                "/graphql",
-                method="POST",
-                json=data,
-                paginator=paginator,
-                data_selector="data",
+                resource_name="repositories_graphql",
+                scope_name="organization",
+                scope_value=org_name,
+                log=logger,
             ):
-                repos_page = page_data[0]["organization"]["repositories"]
+                repos_page = page_data["organization"]["repositories"]
                 for repo in repos_page["nodes"]:
                     repo_record = {**repo}
                     branch_rulesets = repo_record.pop("branchRulesets", None) or {}
@@ -952,22 +956,29 @@ def repositories_graphql(ctx: SourceContext):
                 if isinstance(page_info, dict):
                     repository_cursor = page_info.get("endCursor")
         except Exception as e:
+            failure = e.error if isinstance(e, AdaptiveGraphQLPageError) else e
+            failure_cursor = (
+                e.cursor if isinstance(e, AdaptiveGraphQLPageError) else repository_cursor
+            )
+            page_size = e.page_size if isinstance(e, AdaptiveGraphQLPageError) else None
             logger.error(
                 "Error in resource 'repositories_graphql' processing organization '%s' "
-                "at repository cursor %r after emitting %d repositories "
+                "at repository cursor %r with page size %r after emitting %d repositories "
                 "(%s): %s",
                 org_name,
-                repository_cursor,
+                failure_cursor,
+                page_size,
                 emitted_repositories,
-                type(e).__name__,
-                e,
+                type(failure).__name__,
+                failure,
                 extra={
                     "resource": "repositories_graphql",
                     "phase": "resource_iteration",
                     "org_name": org_name,
-                    "repository_cursor": repository_cursor,
+                    "repository_cursor": failure_cursor,
+                    "page_size": page_size,
                     "emitted_repositories": emitted_repositories,
-                    "error_type": type(e).__name__,
+                    "error_type": type(failure).__name__,
                 },
             )
             continue

{openhound_github-0.6.4 → openhound_github-0.6.6}/src/openhound_github/source.py

@@ -53,6 +53,7 @@ class SourceContext:
     github_deployment_id: str = DEFAULT_GITHUB_DEPLOYMENT_ID
     github_web_origin: str = DEFAULT_GITHUB_WEB_ORIGIN
     cache_lock: Lock = field(default_factory=Lock)
+    organizations_cache: dict[str, dict[str, Any]] = field(default_factory=dict)
     app_cache: dict[str, dict[str, Any]] = field(default_factory=dict)
     actions_permissions_cache: dict[str, dict[str, Any]] = field(default_factory=dict)
     runner_permissions_cache: dict[str, dict[str, Any]] = field(default_factory=dict)
@@ -63,6 +64,31 @@ class SourceContext:
         return [org.org_name for org in self.organizations or []]
 
 
+def _canonicalize_org_names(ctx: SourceContext) -> None:
+    for org in ctx.organizations or []:
+        configured_name = org.org_name
+        try:
+            org_data = org.client.get(f"/orgs/{configured_name}").json()
+        except Exception as exc:
+            logger.warning(
+                "Unable to resolve canonical GitHub login for organization '%s': %s",
+                configured_name,
+                exc,
+            )
+            continue
+
+        canonical_name = org_data.get("login")
+        if not isinstance(canonical_name, str) or not canonical_name:
+            logger.warning(
+                "GitHub organization response for '%s' did not include a canonical login",
+                configured_name,
+            )
+            continue
+
+        org.org_name = canonical_name
+        ctx.organizations_cache[canonical_name] = org_data
+
+
 @configspec
 class GithubCredentials(CredentialsConfiguration):
     org_name: str | None = None
@@ -228,6 +254,7 @@ def source(
             )
         )
 
+        _canonicalize_org_names(ctx)
         return organization_resources(ctx)
 
     else:
@@ -255,4 +282,5 @@ def source(
                 github_web_origin=github_web_origin,
             )
         )
+        _canonicalize_org_names(ctx)
         return organization_resources(ctx)