evalstate/diffusers-pr-api

1from __future__ import annotations2 3import os4import subprocess5from dataclasses import dataclass6from pathlib import Path7from typing import Any8 9 10def _read_gh_token() -> str | None:11    try:12        result = subprocess.run(13            ["gh", "auth", "token"],14            check=True,15            capture_output=True,16            text=True,17        )18    except (OSError, subprocess.CalledProcessError):19        return None20    token = result.stdout.strip()21    return token or None22 23 24def _read_dotenv_token() -> str | None:25    for directory in (Path.cwd(), *Path.cwd().parents):26        path = directory / ".env"27        if not path.exists():28            continue29        values: dict[str, str] = {}30        for line in path.read_text(encoding="utf-8").splitlines():31            line = line.strip()32            if not line or line.startswith("#") or "=" not in line:33                continue34            key, value = line.split("=", 1)35            values[key.strip()] = value.strip().strip("'").strip('"')36        for key in ("GITHUB_TOKEN", "GRAPHQL_TOKEN", "GH_TOKEN"):37            token = values.get(key)38            if token:39                return token40    return None41 42 43def resolve_github_token() -> str | None:44    return (45        os.environ.get("GITHUB_TOKEN")46        or os.environ.get("GRAPHQL_TOKEN")47        or os.environ.get("GH_TOKEN")48        or _read_dotenv_token()49        or _read_gh_token()50    )51 52 53@dataclass(slots=True)54class RepoRef:55    owner: str56    name: str57 58    @classmethod59    def parse(cls, raw: str) -> RepoRef:60        owner, sep, name = raw.partition("/")61        if not sep or not owner or not name:62            raise ValueError(f"Expected REPO in owner/name form, got: {raw!r}")63        return cls(owner=owner, name=name)64 65    @property66    def slug(self) -> str:67        return f"{self.owner}/{self.name}"68 69 70@dataclass(slots=True)71class PipelineOptions:72    repo: RepoRef73    output_dir: Path74    since: str | None75    resume: bool76    http_timeout: int77    http_max_retries: int78    max_issues: int | None79    max_prs: int | None80    max_issue_comments: int | None81    max_reviews_per_pr: int | None82    max_review_comments_per_pr: int | None83    fetch_timeline: bool84    new_contributor_report: bool85    new_contributor_window_days: int86    new_contributor_max_authors: int87    issue_max_age_days: int | None88    pr_max_age_days: int | None89 90 91@dataclass(slots=True)92class AnalysisOptions:93    snapshot_dir: Path | None94    output_dir: Path95    output: Path | None96    hf_repo_id: str | None97    hf_revision: str | None98    hf_materialize_dir: Path | None99    ranking_backend: str100    model: str101    max_clusters: int102    hybrid_llm_concurrency: int = 1103    open_prs_only: bool = False104    cached_analysis: bool = False105    pr_template_cleanup_mode: str = "merge_defaults"106    pr_template_strip_html_comments: bool = True107    pr_template_trim_closing_reference_prefix: bool = True108    pr_template_section_patterns: tuple[str, ...] = ()109    pr_template_line_patterns: tuple[str, ...] = ()110    cluster_suppression_rules: tuple[dict[str, Any], ...] = ()111 112    def __post_init__(self) -> None:113        if self.hybrid_llm_concurrency < 1:114            raise ValueError("hybrid_llm_concurrency must be >= 1")115 116 117@dataclass(slots=True)118class MarkdownReportOptions:119    input: Path120    output: Path | None121    snapshot_dir: Path | None122 123 124@dataclass(slots=True)125class NewContributorReportOptions:126    snapshot_dir: Path | None127    output_dir: Path128    output: Path | None129    json_output: Path | None130    window_days: int131    max_authors: int132    hf_repo_id: str | None = None133    hf_revision: str | None = None134    hf_materialize_dir: Path | None = None135 136 137@dataclass(slots=True)138class DashboardDataOptions:139    snapshot_dir: Path | None140    output_dir: Path141    analysis_input: Path | None142    contributors_input: Path | None143    pr_scope_input: Path | None144    window_days: int145    hf_repo_id: str | None = None146    hf_revision: str | None = None147    hf_materialize_dir: Path | None = None148    snapshot_root: Path | None = None149 150 151@dataclass(slots=True)152class DeployDashboardOptions:153    pipeline_data_dir: Path154    web_dir: Path155    snapshot_dir: Path | None156    analysis_input: Path | None157    contributors_input: Path | None158    pr_scope_input: Path | None159    hf_repo_id: str | None160    hf_revision: str | None161    hf_materialize_dir: Path | None162    refresh_contributors: bool163    dashboard_window_days: int164    contributor_window_days: int165    contributor_max_authors: int166    private_space: bool167    commit_message: str168    space_id: str169    space_title: str | None170    space_emoji: str171    space_color_from: str172    space_color_to: str173    space_short_description: str174    dataset_id: str | None175    space_tags: str | None176 177 178@dataclass(slots=True)179class PrScopeOptions:180    snapshot_dir: Path | None181    output_dir: Path182    output: Path | None183    hf_repo_id: str | None184    hf_revision: str | None185    hf_materialize_dir: Path | None186    cluster_suppression_rules: tuple[dict[str, Any], ...] = ()187 188 189@dataclass(slots=True)190class PrSearchRefreshOptions:191    snapshot_dir: Path | None192    output_dir: Path193    db: Path | None194    hf_repo_id: str | None195    hf_revision: str | None196    hf_materialize_dir: Path | None197    include_drafts: bool = False198    include_closed: bool = False199    limit_prs: int | None = None200    replace_active: bool = True201    cluster_suppression_rules: tuple[dict[str, Any], ...] = ()202 203 204@dataclass(slots=True)205class CheckpointImportOptions:206    source_repo_id: str207    output_dir: Path208    checkpoint_id: str | None209    checkpoint_root: str | None210    publish_repo_id: str | None211    private_hf_repo: bool212    force: bool213 214 215@dataclass(slots=True)216class SnapshotAdoptOptions:217    snapshot_dir: Path218    output_dir: Path219    next_since: str | None220 221 222@dataclass(slots=True)223class DatasetRefreshOptions:224    repo: RepoRef225    hf_repo_id: str226    private_hf_repo: bool227    max_issues: int | None228    max_prs: int | None229    max_issue_comments: int | None230    max_reviews_per_pr: int | None231    max_review_comments_per_pr: int | None232    fetch_timeline: bool233    new_contributor_report: bool234    new_contributor_window_days: int235    new_contributor_max_authors: int236    http_timeout: int237    http_max_retries: int238    checkpoint_every_comments: int239    checkpoint_every_prs: int240    cluster_suppression_rules: tuple[dict[str, Any], ...] = ()241 242 243@dataclass(slots=True)244class PublishAnalysisArtifactsOptions:245    output_dir: Path246    snapshot_dir: Path | None247    analysis_input: Path | None248    hf_repo_id: str249    analysis_id: str250    canonical: bool = False251    save_cache: bool = False252    private_hf_repo: bool = False253 254 255@dataclass(slots=True)256class SaveCacheOptions:257    output_dir: Path258    snapshot_dir: Path | None259    hf_repo_id: str260    private_hf_repo: bool = False261 262 263@dataclass(slots=True)264class DatasetStatusOptions:265    output_dir: Path266    hf_repo_id: str | None267    hf_revision: str | None268    repo: str | None = None269    json_output: bool = False270