Skip to content

Local UI to run and train LLMs and diffusion models. Supports GGUF, MLX, Qwen3.8, DeepSeek-V4.1, MiniMax-H3, Gemma 4, FLUX and more.

12.3 KB
Raw Blame
# SPDX-License-Identifier: AGPL-3.0-only
# Copyright 2026-present the Unsloth AI Inc. team. All rights reserved. See /studio/LICENSE.AGPL-3.0

"""New carried changes between two published llama.cpp prebuilts. The release body is cumulative, so the banner must diff the installed and target bodies; showing the target body alone relabels old carried PRs as new."""

from __future__ import annotations

import http.client
import json
import os
import re
import time
import urllib.error
import urllib.parse
import urllib.request
from typing import Optional

import structlog

from utils.auth_safe import auth_safe_open
from utils.prebuilt.freshness_flow import (
    RELEASE_CACHE_TTL_SECONDS,
    RELEASE_FAILURE_CACHE_TTL_SECONDS,
    github_rate_limit_remaining,
    hold_github_api,
    rate_limit_wait,
)

logger = structlog.get_logger(__name__)

MAX_CHANGES = 50
# The only repo whose notes this module can read: generated, cumulative, one bullet per carried PR. --published-repo can point elsewhere, and a per-release body says nothing about what is still carried.
CUMULATIVE_NOTES_REPO = "unslothai/llama.cpp"
# A release body is a few KB; the cap only bounds a far side that misbehaves.
MAX_RELEASE_BYTES = 4 * 1024 * 1024
# Without a floor, a held-down Retry is two uncached GitHub calls per click.
FORCE_REFRESH_MIN_INTERVAL_SECONDS = 30.0
_REPO = re.compile(r"^[A-Za-z0-9_.-]+/[A-Za-z0-9_.-]+$")
# "." is in _REPO's class, so "owner/.." walks out of /repos/ on a normalizing proxy.
_DOT_SEGMENT = re.compile(r"^\.+$")
_BULLET = re.compile(r"^ {0,3}[-*+]\s+(.+?)\s*$")
_LINK = re.compile(r"\[([^\]]+)]\((https://github\.com/[^\s)]+)\)")
_PR_URL = re.compile(r"^https://github\.com/([^/]+/[^/]+)/pull/(\d+)(?:/|$)", re.I)
_ISSUE_URL = re.compile(r"^https://github\.com/([^/]+/[^/]+)/issues/(\d+)(?:/|$)", re.I)
_TEXT_REFERENCE = re.compile(r"(?<![\w./-])([A-Za-z0-9_.-]+(?:/[A-Za-z0-9_.-]+)?)#(\d+)\b")

_release_memo: dict[tuple[str, str], tuple[float, dict]] = {}
_release_failed_at: dict[tuple[str, str], float] = {}
_release_forced_at: dict[tuple[str, str], float] = {}


def _valid_repo(repo: str) -> bool:
    """``owner/name``, with no segment that is only dots."""
    if not isinstance(repo, str) or not _REPO.fullmatch(repo):
        return False
    return not any(_DOT_SEGMENT.fullmatch(part) for part in repo.split("/"))


def _is_cumulative_repo(repo: str) -> bool:
    """Case-folded: GitHub owner/name is case-insensitive and --published-repo persists whatever spelling was typed."""
    return repo.casefold() == CUMULATIVE_NOTES_REPO.casefold()


def _fetch_release(
    repo: str,
    tag: str,
    timeout: float = 5.0,
) -> Optional[dict]:
    """One exact GitHub release. None on invalid input or any failure."""
    if not _valid_repo(repo) or not tag:
        return None
    from utils.utils import call_with_deadline

    try:
        return call_with_deadline(
            lambda: _fetch_release_blocking(repo, tag, timeout),
            timeout + 1,
            name = "llama-changelog-fetch",
        )
    except TimeoutError as exc:
        logger.debug("llama changelog fetch failed", repo = repo, tag = tag, error = str(exc))
        return None


def _fetch_release_blocking(repo: str, tag: str, timeout: float) -> Optional[dict]:
    if github_rate_limit_remaining() > 0:
        return None
    encoded_tag = urllib.parse.quote(tag, safe = "")
    headers = {
        "Accept": "application/vnd.github+json",
        "User-Agent": "unsloth-studio-llama-changelog",
    }
    token = os.environ.get("GITHUB_TOKEN") or os.environ.get("GH_TOKEN")
    if token:
        headers["Authorization"] = f"Bearer {token}"
    request = urllib.request.Request(
        f"https://api.github.com/repos/{repo}/releases/tags/{encoded_tag}",
        headers = headers,
    )
    try:
        with auth_safe_open(request, timeout = timeout) as response:
            # One byte past the cap: reject an oversized body without buffering it.
            raw = response.read(MAX_RELEASE_BYTES + 1)
        if len(raw) > MAX_RELEASE_BYTES:
            logger.debug("llama changelog release too large", repo = repo, tag = tag)
            return None
        payload = json.loads(raw.decode("utf-8"))
    except urllib.error.HTTPError as exc:
        wait = rate_limit_wait(exc)
        hold_github_api(wait)
        logger.debug(
            "llama changelog fetch failed", repo = repo, tag = tag, error = str(exc), backoff_seconds = wait
        )
        return None
    except (
        urllib.error.URLError,
        OSError,
        # A truncated read raises HTTPException, which is not an OSError.
        http.client.HTTPException,
        UnicodeDecodeError,
        json.JSONDecodeError,
    ) as exc:
        logger.debug("llama changelog fetch failed", repo = repo, tag = tag, error = str(exc))
        return None
    return payload if isinstance(payload, dict) else None


def _release_for_tag(
    repo: str,
    tag: str,
    *,
    force_refresh: bool = False,
) -> Optional[dict]:
    """Exact release with 24h success and 60s failure memoization."""
    key = (repo, tag)
    # Memory-only, so monotonic throughout: a backward clock step must not be able to extend the TTL. freshness_flow uses wall time because it persists to disk.
    now = time.monotonic()
    # Retrying into a rate limit only delays the reset; checked before the debounce so the slot is not burnt.
    if github_rate_limit_remaining() > 0:
        force_refresh = False
    if force_refresh:
        forced_at = _release_forced_at.get(key)
        if forced_at is not None and now - forced_at < FORCE_REFRESH_MIN_INTERVAL_SECONDS:
            force_refresh = False
        else:
            _release_forced_at[key] = now
    if not force_refresh:
        failed_at = _release_failed_at.get(key)
        cached = _release_memo.get(key)
        fresh = cached is not None and now - cached[0] < RELEASE_CACHE_TTL_SECONDS
        if failed_at is not None and now - failed_at < RELEASE_FAILURE_CACHE_TTL_SECONDS:
            # Suppress the retry, but never resurrect an entry past its TTL.
            return cached[1] if fresh else None
        if fresh:
            return cached[1]
    release = _fetch_release(repo, tag)
    if release is None:
        _release_failed_at[key] = time.monotonic()
        # Last-good fallback only within the TTL, or an unreachable release keeps answering stale and the panel presents that as matched.
        cached = _release_memo.get(key)
        if cached and time.monotonic() - cached[0] < RELEASE_CACHE_TTL_SECONDS:
            return cached[1]
        return None
    _release_failed_at.pop(key, None)
    _release_memo[key] = (time.monotonic(), release)
    return release


def _plain_text(markdown: str) -> str:
    text = _LINK.sub(lambda match: match.group(1), markdown)
    # Underscores in ROCm_Host / GGML_CUDA_ENABLE_UNIFIED_MEMORY are text, not emphasis.
    text = text.replace("`", "").replace("**", "")
    return re.sub(r"\s+", " ", text).strip()


def _entry(markdown: str) -> dict:
    # Metadata starts at " ([", so a title keeps its parens: GLM-5-Next (GLM-5.3-Flash). rfind, not find: metadata is the LAST parenthesised group, and a title may contain " ([" as in "vulkan: handle ([a],[b]) tuples ([#5](...))".
    metadata_at = markdown.rfind(" ([")
    summary_markdown = markdown[:metadata_at] if metadata_at >= 0 else markdown
    links = []
    for label, url in _LINK.findall(markdown[metadata_at:] if metadata_at >= 0 else ""):
        clean_label = _plain_text(label)
        if "/commits/" in url and not clean_label.lower().startswith("commit"):
            clean_label = f"commit {clean_label}"
        links.append({"label": clean_label, "url": url})
    return {"summary": _plain_text(summary_markdown), "links": links}


def _identities(markdown: str) -> set[str]:
    """Stable aliases for one carried change: a patch migrated to an Unsloth carry PR links that PR but still says ``ggml-org#24423``, and both must match."""
    identities = set()
    # One namespace: GitHub numbers issues and PRs together, so ``/issues/900``, ``/pull/900`` and ``repo#900`` are the same object. Separate prefixes only miss.
    for _label, url in _LINK.findall(markdown):
        match = _PR_URL.match(url) or _ISSUE_URL.match(url)
        if match:
            identities.add(f"ref:{match.group(1).lower()}#{match.group(2)}")
    for repo, number in _TEXT_REFERENCE.findall(_plain_text(markdown)):
        # Shorthand omits the repo: ``ggml-org#24423`` is ``ggml-org/llama.cpp#24423``.
        if "/" not in repo:
            repo = f"{repo}/llama.cpp"
        identities.add(f"ref:{repo.lower()}#{number}")
    if not identities:
        identities.add(f"text:{_entry(markdown)['summary'].casefold()}")
    return identities


def _bullets(body: object) -> list[str]:
    if not isinstance(body, str):
        return []
    return [match.group(1) for line in body.splitlines() if (match := _BULLET.match(line))]


def release_page_url(repo: str, tag: str) -> Optional[str]:
    """The human release page, so a failed comparison can still offer a way to read the notes on GitHub. None when the repo is not a safe ``owner/name``."""
    if not _valid_repo(repo) or not tag:
        return None
    return f"https://github.com/{repo}/releases/tag/{urllib.parse.quote(tag, safe = '')}"


def unavailable_reason(repo: str, installed_tag: str, latest_tag: str) -> str:
    """Why a comparison could not be made, for a caller holding ``None``. ``notes_not_itemised`` (predates the bullet format) and ``notes_not_comparable`` (non-cumulative repo) are permanent; ``release_notes_unavailable`` may succeed later, so it keeps its Retry."""
    if not _valid_repo(repo) or not installed_tag or not latest_tag:
        return "release_notes_unavailable"
    if not _is_cumulative_repo(repo):
        return "notes_not_comparable"
    # Memoized, so this re-read costs nothing after the comparison's own lookups.
    installed = _release_for_tag(repo, installed_tag)
    if installed is None:
        return "release_notes_unavailable"
    # Only the INSTALLED side is permanent: it shipped before the bullet format and will never gain one. A bad target is the newest release, so it may yet be fixed.
    if not _bullets(installed.get("body")):
        return "notes_not_itemised"
    return "release_notes_unavailable"


def changelog_for_update(
    repo: str,
    installed_tag: str,
    latest_tag: str,
    *,
    force_refresh: bool = False,
) -> Optional[dict]:
    """Return only target bullets absent from the installed release. None means no comparison was possible; do not fall back to the cumulative body."""
    if not repo or not installed_tag or not latest_tag or installed_tag == latest_tag:
        return None
    if not _is_cumulative_repo(repo):
        return None
    installed = _release_for_tag(repo, installed_tag, force_refresh = force_refresh)
    latest = _release_for_tag(repo, latest_tag, force_refresh = force_refresh)
    if installed is None or latest is None:
        return None

    # Releases before b9625-mix-2d6bd50 (2026-06-14) name carries in prose, so no bullets means unknown, not "carries nothing".
    installed_bullets = _bullets(installed.get("body"))
    if not installed_bullets:
        return None
    # A prose-only target says "carries nothing"; a missing or blank one says nothing at all, and "no new changes" would claim a comparison never made.
    latest_body = latest.get("body")
    if not isinstance(latest_body, str) or not latest_body.strip():
        return None
    latest_bullets = _bullets(latest_body)
    old_identities = set().union(*(_identities(item) for item in installed_bullets))
    new_items = []
    seen = set()
    for item in latest_bullets:
        identities = _identities(item)
        if identities & old_identities or identities & seen:
            continue
        parsed = _entry(item)
        if not parsed["summary"]:
            # A bullet that is never shown must not suppress a later one via `seen`.
            continue
        seen.update(identities)
        new_items.append(parsed)

    total = len(new_items)
    release_url = latest.get("html_url")
    if not isinstance(release_url, str) or not release_url.startswith("https://github.com/"):
        encoded_tag = urllib.parse.quote(latest_tag, safe = "")
        release_url = f"https://github.com/{repo}/releases/tag/{encoded_tag}"
    return {
        "changes": new_items[:MAX_CHANGES],
        "total_changes": total,
        "truncated": total > MAX_CHANGES,
        "release_url": release_url,
    }