Improve caching by comparing file hashes as fallback for mtime and size

CHANGES.md

@@ -31,6 +31,7 @@
 <!-- Changes that improve Black's performance. -->
 
 - Avoid importing `IPython` if notebook cells do not contain magics (#3782)
+- Improve caching by comparing file hashes as fallback for mtime and size. (#3821)
 
 ### Output
 

docs/contributing/reference/reference_classes.rst

@@ -186,6 +186,13 @@ Black Classes
     :show-inheritance:
     :members:
 
+:class:`Cache`
+------------------------
+
+.. autoclass:: black.cache.Cache
+    :show-inheritance:
+    :member:
+
 Enum Classes
 ~~~~~~~~~~~~~
 

docs/contributing/reference/reference_functions.rst

@@ -94,18 +94,10 @@ Split functions
 Caching
 -------
 
-.. autofunction:: black.cache.filter_cached
-
 .. autofunction:: black.cache.get_cache_dir
 
 .. autofunction:: black.cache.get_cache_file
 
-.. autofunction:: black.cache.get_cache_info
-
-.. autofunction:: black.cache.read_cache
-
-.. autofunction:: black.cache.write_cache
-
 Utilities
 ---------
 

pyproject.toml

@@ -67,7 +67,7 @@ dependencies = [
   "pathspec>=0.9.0",
   "platformdirs>=2",
   "tomli>=1.1.0; python_version < '3.11'",
-  "typing_extensions>=3.10.0.0; python_version < '3.10'",
+  "typing_extensions>=4.0.1; python_version < '3.11'",
 ]
 dynamic = ["readme", "version"]
 

src/black/__init__.py

@@ -34,7 +34,7 @@
 from pathspec.patterns.gitwildmatch import GitWildMatchPatternError
 
 from _black_version import version as __version__
-from black.cache import Cache, get_cache_info, read_cache, write_cache
+from black.cache import Cache
 from black.comments import normalize_fmt_off
 from black.const import (
     DEFAULT_EXCLUDES,
@@ -775,12 +775,9 @@ def reformat_one(
             if format_stdin_to_stdout(fast=fast, write_back=write_back, mode=mode):
                 changed = Changed.YES
         else:
-            cache: Cache = {}
+            cache = Cache.read(mode)
             if write_back not in (WriteBack.DIFF, WriteBack.COLOR_DIFF):
-                cache = read_cache(mode)
-                res_src = src.resolve()
-                res_src_s = str(res_src)
-                if res_src_s in cache and cache[res_src_s] == get_cache_info(res_src):
+                if not cache.is_changed(src):
                     changed = Changed.CACHED
             if changed is not Changed.CACHED and format_file_in_place(
                 src, fast=fast, write_back=write_back, mode=mode
@@ -789,7 +786,8 @@ def reformat_one(
             if (write_back is WriteBack.YES and changed is not Changed.CACHED) or (
                 write_back is WriteBack.CHECK and changed is Changed.NO
             ):
-                write_cache(cache, [src], mode)
+                cache.update([src])
+                cache.write()
         report.done(src, changed)
     except Exception as exc:
         if report.verbose:

src/black/cache.py

@@ -1,21 +1,28 @@
 """Caching of formatted files with feature-based invalidation."""
-
+import hashlib
 import os
 import pickle
+import sys
 import tempfile
+from dataclasses import dataclass, field
 from pathlib import Path
-from typing import Dict, Iterable, Set, Tuple
+from typing import Dict, Iterable, NamedTuple, Set, Tuple
 
 from platformdirs import user_cache_dir
 
 from _black_version import version as __version__
 from black.mode import Mode
 
-# types
-Timestamp = float
-FileSize = int
-CacheInfo = Tuple[Timestamp, FileSize]
-Cache = Dict[str, CacheInfo]
+if sys.version_info >= (3, 11):
+    from typing import Self
+else:
+    from typing_extensions import Self
+
+
+class FileData(NamedTuple):
+    st_mtime: float
+    st_size: int
+    hash: str
 
 
 def get_cache_dir() -> Path:
@@ -37,61 +44,103 @@ def get_cache_dir() -> Path:
 CACHE_DIR = get_cache_dir()
 
 
-def read_cache(mode: Mode) -> Cache:
-    """Read the cache if it exists and is well formed.
-
-    If it is not well formed, the call to write_cache later should resolve the issue.
-    """
-    cache_file = get_cache_file(mode)
-    if not cache_file.exists():
-        return {}
-
-    with cache_file.open("rb") as fobj:
-        try:
-            cache: Cache = pickle.load(fobj)
-        except (pickle.UnpicklingError, ValueError, IndexError):
-            return {}
-
-    return cache
-
-
 def get_cache_file(mode: Mode) -> Path:
     return CACHE_DIR / f"cache.{mode.get_cache_key()}.pickle"
 
 
-def get_cache_info(path: Path) -> CacheInfo:
-    """Return the information used to check if a file is already formatted or not."""
-    stat = path.stat()
-    return stat.st_mtime, stat.st_size
-
+@dataclass
+class Cache:
+    mode: Mode
+    cache_file: Path
+    file_data: Dict[str, FileData] = field(default_factory=dict)
+
+    @classmethod
+    def read(cls, mode: Mode) -> Self:
+        """Read the cache if it exists and is well formed.
+
+        If it is not well formed, the call to write later should
+        resolve the issue.
+        """
+        cache_file = get_cache_file(mode)
+        if not cache_file.exists():
+            return cls(mode, cache_file)
+
+        with cache_file.open("rb") as fobj:
+            try:
+                file_data: Dict[str, FileData] = pickle.load(fobj)
+            except (pickle.UnpicklingError, ValueError, IndexError):
+                return cls(mode, cache_file)
+
+        return cls(mode, cache_file, file_data)
+
+    @staticmethod
+    def hash_digest(path: Path) -> str:
+        """Return hash digest for path."""
+
+        with open(path, "rb") as fp:
+            data = fp.read()
+        return hashlib.sha256(data).hexdigest()
+
+    @staticmethod
+    def get_file_data(path: Path) -> FileData:
+        """Return file data for path."""
+
+        stat = path.stat()
+        hash = Cache.hash_digest(path)
+        return FileData(stat.st_mtime, stat.st_size, hash)
+
+    def is_changed(self, source: Path) -> bool:
+        """Check if source has changed compared to cached version."""
+        res_src = source.resolve()
+        old = self.file_data.get(str(res_src))
+        if old is None:
+            return True
+
+        st = res_src.stat()
+        if st.st_size != old.st_size or int(st.st_mtime) != int(old.st_mtime):
+            new_hash = Cache.hash_digest(res_src)
+            if st.st_size != old.st_size or new_hash != old.hash:
+                return True
+        return False
+
+    def filtered_cached(self, sources: Iterable[Path]) -> Tuple[Set[Path], Set[Path]]:
+        """Split an iterable of paths in `sources` into two sets.
+
+        The first contains paths of files that modified on disk or are not in the
+        cache. The other contains paths to non-modified files.
+        """
+        changed: Set[Path] = set()
+        done: Set[Path] = set()
+        for src in sources:
+            res_src = src.resolve()
+            old = self.file_data.get(str(res_src))
+            st = res_src.stat()
+
+            if old is None:
+                changed.add(src)
+                continue
+            elif st.st_size != old.st_size or int(st.st_mtime) != int(old.st_mtime):
+                new_hash = Cache.hash_digest(res_src)
+                if st.st_size != old.st_size or new_hash != old.hash:
+                    changed.add(src)
+                    continue
+            done.add(src)
+        return changed, done
 
-def filter_cached(cache: Cache, sources: Iterable[Path]) -> Tuple[Set[Path], Set[Path]]:
-    """Split an iterable of paths in `sources` into two sets.
+    def update(self, sources: Iterable[Path]) -> None:
+        """Update the cache file data."""
+        self.file_data.update(
+            **{str(src.resolve()): Cache.get_file_data(src) for src in sources}
+        )
 
-    The first contains paths of files that modified on disk or are not in the
-    cache. The other contains paths to non-modified files.
-    """
-    todo, done = set(), set()
-    for src in sources:
-        res_src = src.resolve()
-        if cache.get(str(res_src)) != get_cache_info(res_src):
-            todo.add(src)
-        else:
-            done.add(src)
-    return todo, done
-
-
-def write_cache(cache: Cache, sources: Iterable[Path], mode: Mode) -> None:
-    """Update the cache file."""
-    cache_file = get_cache_file(mode)
-    try:
-        CACHE_DIR.mkdir(parents=True, exist_ok=True)
-        new_cache = {
-            **cache,
-            **{str(src.resolve()): get_cache_info(src) for src in sources},
-        }
-        with tempfile.NamedTemporaryFile(dir=str(cache_file.parent), delete=False) as f:
-            pickle.dump(new_cache, f, protocol=4)
-        os.replace(f.name, cache_file)
-    except OSError:
-        pass
+    def write(self) -> None:
+        """Write a new cache file."""
+        try:
+            CACHE_DIR.mkdir(parents=True, exist_ok=True)
+            with tempfile.NamedTemporaryFile(
+                dir=str(self.cache_file.parent), delete=False
+            ) as f:
+                pickle.dump(self.file_data, f, protocol=4)
+            os.replace(f.name, self.cache_file)
+        except OSError:
+            pass

src/black/concurrency.py

@@ -17,7 +17,7 @@
 from mypy_extensions import mypyc_attr
 
 from black import WriteBack, format_file_in_place
-from black.cache import Cache, filter_cached, read_cache, write_cache
+from black.cache import Cache
 from black.mode import Mode
 from black.output import err
 from black.report import Changed, Report
@@ -133,10 +133,9 @@ async def schedule_formatting(
     `write_back`, `fast`, and `mode` options are passed to
     :func:`format_file_in_place`.
     """
-    cache: Cache = {}
+    cache = Cache.read(mode)
     if write_back not in (WriteBack.DIFF, WriteBack.COLOR_DIFF):
-        cache = read_cache(mode)
-        sources, cached = filter_cached(cache, sources)
+        sources, cached = cache.filtered_cached(sources)
         for src in sorted(cached):
             report.done(src, Changed.CACHED)
     if not sources:
@@ -185,4 +184,5 @@ async def schedule_formatting(
     if cancelled:
         await asyncio.gather(*cancelled, return_exceptions=True)
     if sources_to_cache:
-        write_cache(cache, sources_to_cache, mode)
+        cache.update(sources_to_cache)
+        cache.write()