Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Auto track binary files #828

Merged
merged 9 commits into from Apr 21, 2022
Merged

Auto track binary files #828

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
a4e9b19
Auto track binary files
LysandreJik Apr 8, 2022
35b480c
Add null character check
LysandreJik Apr 8, 2022
57f8613
Tests
LysandreJik Apr 8, 2022
b8bbe4f
Apply suggestions from code review
LysandreJik Apr 12, 2022
17a1331
Only read 512 bytes
LysandreJik Apr 12, 2022
5f48d6c
Address Eliott's comments
LysandreJik Apr 12, 2022
6d27a15
Split test
LysandreJik Apr 12, 2022
9fe6bb6
Address review comments
LysandreJik Apr 18, 2022
cbfdce5
10MB maximum
LysandreJik Apr 21, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
20 changes: 12 additions & 8 deletions src/huggingface_hub/hf_api.py
View file
Open in desktop
Expand Up @@ -569,19 +569,23 @@ def _validate_or_retrieve_token(
function_name: Optional[str] = None,
):
"""
Retrieves and validates stored token or validates passed token.
Args:
Retrieves and validates stored token or validates passed token.
token (``str``, `optional`):
Hugging Face token. Will default to the locally saved token if not provided.
Hugging Face token. Will default to the locally saved token if
not provided.
name (``str``, `optional`):
Name of the repository. This is deprecated in favor of repo_id and will be removed in v0.7.
Name of the repository. This is deprecated in favor of repo_id
and will be removed in v0.7.
function_name (``str``, `optional`):
If _validate_or_retrieve_token is called from a function, name of that function to be passed inside deprecation warning.
If _validate_or_retrieve_token is called from a function, name
of that function to be passed inside deprecation warning.
Returns:
Validated token and the name of the repository.
Raises:
:class:`EnvironmentError`: If the token is not passed and there's no token saved locally.
:class:`ValueError`: If organization token or invalid token is passed.
:class:`EnvironmentError`: If the token is not passed and there's no
token saved locally. :class:`ValueError`: If organization token or
invalid token is passed.
"""
if token is None or token is True:
token = HfFolder.get_token()
Expand Down Expand Up @@ -1847,8 +1851,8 @@ def get_token(cls) -> Optional[str]:
"""
Get token or None if not existent.

Note that a token can be also provided using the `HUGGING_FACE_HUB_TOKEN`
environment variable.
Note that a token can be also provided using the
`HUGGING_FACE_HUB_TOKEN` environment variable.

Returns:
`str` or `None`: The token, `None` if it doesn't exist.
Expand Down
71 changes: 66 additions & 5 deletions src/huggingface_hub/repository.py
View file
Open in desktop
Expand Up @@ -226,6 +226,27 @@ def is_git_ignored(filename: Union[str, Path]) -> bool:
return is_ignored


def is_binary_file(filename: Union[str, Path]) -> bool:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Other than reading the first few bytes, I think this answer gives a more accurate way of checking if the file is binary or not: https://stackoverflow.com/a/7392391/2536294

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I implemented what you propose in 9fe6bb6, let me know if that's what you had in mind. Cc @coyotte508

"""
Check if file is a binary file.

Args:
filename (`str` or `Path`):
The filename to check.

Returns:
`bool`: `True` if the file passed is a binary file, `False` otherwise.
"""
try:
with open(filename) as f:
content = f.read()
LysandreJik marked this conversation as resolved.
Show resolved Hide resolved
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
content = f.read()
content = f.read(1024) # or 512 if we want to be consistent with the backend

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This was discussed above here: #828 (comment)

Do you disagree with the conclusion?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Didn't see that. Yeah I don't think we should ever read 11GB of data into memory. This will most certainly crash most people's systems. I'd be happy if we read like 10MB instead of 1MB to reduce the probability of a false detection, which should address those concerns. If we really do want to read a lot more, we should read in chunks. As python's docs state:

To read a file’s contents, call f.read(size), which reads some quantity of data and returns it as a string (in text mode) or bytes object (in binary mode). size is an optional numeric argument. When size is omitted or negative, the entire contents of the file will be read and returned; it’s your problem if the file is twice as large as your machine’s memory.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The solution currently implemented loads a maximum of 10MB in memory when calling git_add: it tracks large files before tracking binary files.

When tracking binary files, it looks at files which are not yet tracked with lfs, eliminating large files.

Instead of the 1MB limit that you propose here, we could instead put a max of 10MB here, which will only be triggered when auto_track_binary_files is called independently of git_add (which is a possibility!).

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yes, but you also want to have this method public, which means people can call it before having tracked large files.

I'm happy with your suggestion of 10MB.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sounds good, thanks for your review. Addressed in cbfdce5


# Check for the presence of the null character in the string
return "\x00" in content
except UnicodeDecodeError:
return True


def files_to_be_staged(pattern: str, folder: Union[str, Path]) -> List[str]:
"""
Returns a list of filenames that are to be staged.
Expand Down Expand Up @@ -485,8 +506,8 @@ def __init__(
skip_lfs_files (`bool`, *optional*, defaults to `False`):
whether to skip git-LFS files or not.
client (`HfApi`, *optional*):
Instance of HfApi to use when calling the HF Hub API.
A new instance will be created if this is left to `None`.
Instance of HfApi to use when calling the HF Hub API. A new
instance will be created if this is left to `None`.
"""

os.makedirs(local_dir, exist_ok=True)
Expand Down Expand Up @@ -981,6 +1002,42 @@ def lfs_enable_largefiles(self):
except subprocess.CalledProcessError as exc:
raise EnvironmentError(exc.stderr)

def auto_track_binary_files(self, pattern: Optional[str] = ".") -> List[str]:
adrinjalali marked this conversation as resolved.
Show resolved Hide resolved
"""
Automatically track binary files with git-lfs.

Args:
pattern (`str`, *optional*, defaults to "."):
The pattern with which to track files that are above 10MBs.
LysandreJik marked this conversation as resolved.
Show resolved Hide resolved

Returns:
`List[str]`: List of filenames that are now tracked due to being
binary files
"""
files_to_be_tracked_with_lfs = []

deleted_files = self.list_deleted_files()

for filename in files_to_be_staged(pattern, folder=self.local_dir):
if filename in deleted_files:
continue

path_to_file = os.path.join(os.getcwd(), self.local_dir, filename)
is_binary = is_binary_file(path_to_file)

if (
is_binary
and not is_tracked_with_lfs(path_to_file)
and not is_git_ignored(path_to_file)
):
self.lfs_track(filename)
files_to_be_tracked_with_lfs.append(filename)

# Cleanup the .gitattributes if files were deleted
self.lfs_untrack(deleted_files)

return files_to_be_tracked_with_lfs

def auto_track_large_files(self, pattern: Optional[str] = ".") -> List[str]:
"""
Automatically track large files (files that weigh more than 10MBs) with
Expand Down Expand Up @@ -1090,11 +1147,15 @@ def git_add(
pattern (`str`, *optional*, defaults to "."):
The pattern with which to add files to staging.
auto_lfs_track (`bool`, *optional*, defaults to `False`):
Whether to automatically track large files with git-lfs. Any
file over 10MB in size will be automatically tracked.
Whether to automatically track large and binaryfiles with
LysandreJik marked this conversation as resolved.
Show resolved Hide resolved
git-lfs. Any file over 10MB in size, or in binary format, will
be automatically tracked.
"""
if auto_lfs_track:
tracked_files = self.auto_track_large_files(pattern)
tracked_files = [
*self.auto_track_large_files(pattern),
*self.auto_track_binary_files(pattern),
]
if tracked_files:
logger.warning(
f"Adding files tracked by Git LFS: {tracked_files}. This may take a bit of time if the files are large."
Expand Down
89 changes: 85 additions & 4 deletions tests/test_repository.py
View file
Open in desktop
Expand Up @@ -1075,12 +1075,18 @@ def test_is_tracked_with_lfs_with_pattern(self):
# This content is 20MB (over 10MB)
large_file = [100] * int(4e6)

# This content is binary (contains the null character)
binary_file = "\x00\x00\x00\x00"

with open(f"{WORKING_REPO_DIR}/large_file.txt", "w+") as f:
f.write(json.dumps(large_file))

with open(f"{WORKING_REPO_DIR}/small_file.txt", "w+") as f:
f.write(json.dumps(small_file))

with open(f"{WORKING_REPO_DIR}/binary_file.txt", "w+") as f:
f.write(binary_file)

os.makedirs(f"{WORKING_REPO_DIR}/dir", exist_ok=True)

with open(f"{WORKING_REPO_DIR}/dir/large_file.txt", "w+") as f:
Expand All @@ -1089,20 +1095,30 @@ def test_is_tracked_with_lfs_with_pattern(self):
with open(f"{WORKING_REPO_DIR}/dir/small_file.txt", "w+") as f:
f.write(json.dumps(small_file))

with open(f"{WORKING_REPO_DIR}/dir/binary_file.txt", "w+") as f:
f.write(binary_file)

repo.auto_track_large_files("dir")
repo.auto_track_binary_files("dir")

self.assertFalse(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "large_file.txt"))
)
self.assertFalse(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "small_file.txt"))
)
self.assertFalse(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "binary_file.txt"))
)
self.assertTrue(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "dir/large_file.txt"))
)
self.assertFalse(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "dir/small_file.txt"))
)
self.assertTrue(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "dir/binary_file.txt"))
)

def test_auto_track_large_files(self):
repo = Repository(WORKING_REPO_DIR)
Expand All @@ -1113,35 +1129,48 @@ def test_auto_track_large_files(self):
# This content is 20MB (over 10MB)
large_file = [100] * int(4e6)

# This content is binary (contains the null character)
binary_file = "\x00\x00\x00\x00"
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would consider moving this to another test to keep this test focused in large files (or rename this test, although I think different tests would be better)

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Right, what I was mentioning in the description is that this effectively doubles the time spent on these tests, which is already significant. We're currently sitting at 10+ minutes for the tests, which isn't ideal.

I'm open to moving it around, but would like to emphasize that it will likely end up taking even longer. Let me know if this compromise is okay for you and I'll move it to its own separate test.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good point! Let's keep as is then 😄


with open(f"{WORKING_REPO_DIR}/large_file.txt", "w+") as f:
f.write(json.dumps(large_file))

with open(f"{WORKING_REPO_DIR}/small_file.txt", "w+") as f:
f.write(json.dumps(small_file))

with open(f"{WORKING_REPO_DIR}/binary_file.txt", "w+") as f:
f.write(binary_file)

repo.auto_track_large_files()
repo.auto_track_binary_files()

self.assertTrue(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "large_file.txt"))
)
self.assertFalse(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "small_file.txt"))
)
self.assertTrue(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "binary_file.txt"))
)

def test_auto_track_large_files_ignored_with_gitignore(self):
repo = Repository(WORKING_REPO_DIR)

# This content is 20MB (over 10MB)
large_file = [100] * int(4e6)

# This content is binary (contains the null character)
binary_file = "\x00\x00\x00\x00"

# Test nested gitignores
os.makedirs(f"{WORKING_REPO_DIR}/directory")

with open(f"{WORKING_REPO_DIR}/.gitignore", "w+") as f:
f.write("large_file.txt")
f.write("large_file.txt\nbinary_file.txt")

with open(f"{WORKING_REPO_DIR}/directory/.gitignore", "w+") as f:
f.write("large_file_3.txt")
f.write("large_file_3.txt\nbinary_file_3.txt")

with open(f"{WORKING_REPO_DIR}/large_file.txt", "w+") as f:
f.write(json.dumps(large_file))
Expand All @@ -1157,6 +1186,21 @@ def test_auto_track_large_files_ignored_with_gitignore(self):

repo.auto_track_large_files()

with open(f"{WORKING_REPO_DIR}/binary_file.txt", "w+") as f:
f.write(binary_file)

with open(f"{WORKING_REPO_DIR}/binary_file_2.txt", "w+") as f:
f.write(binary_file)

with open(f"{WORKING_REPO_DIR}/directory/binary_file_3.txt", "w+") as f:
f.write(binary_file)

with open(f"{WORKING_REPO_DIR}/directory/binary_file_4.txt", "w+") as f:
f.write(binary_file)

repo.auto_track_binary_files()

# Large files
self.assertFalse(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "large_file.txt"))
)
Expand All @@ -1175,7 +1219,26 @@ def test_auto_track_large_files_ignored_with_gitignore(self):
)
)

def test_auto_track_large_files_through_git_add(self):
# Binary files
self.assertFalse(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "binary_file.txt"))
)
self.assertTrue(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "binary_file_2.txt"))
)

self.assertFalse(
is_tracked_with_lfs(
os.path.join(WORKING_REPO_DIR, "directory/binary_file_3.txt")
)
)
self.assertTrue(
is_tracked_with_lfs(
os.path.join(WORKING_REPO_DIR, "directory/binary_file_4.txt")
)
)

def test_auto_track_files_through_git_add(self):
repo = Repository(WORKING_REPO_DIR)

# This content is 5MB (under 10MB)
Expand All @@ -1184,12 +1247,18 @@ def test_auto_track_large_files_through_git_add(self):
# This content is 20MB (over 10MB)
large_file = [100] * int(4e6)

# This content is binary (contains the null character)
binary_file = "\x00\x00\x00\x00"

with open(f"{WORKING_REPO_DIR}/large_file.txt", "w+") as f:
f.write(json.dumps(large_file))

with open(f"{WORKING_REPO_DIR}/small_file.txt", "w+") as f:
f.write(json.dumps(small_file))

with open(f"{WORKING_REPO_DIR}/binary_file.txt", "w+") as f:
f.write(binary_file)

repo.git_add(auto_lfs_track=True)

self.assertTrue(
Expand All @@ -1198,8 +1267,11 @@ def test_auto_track_large_files_through_git_add(self):
self.assertFalse(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "small_file.txt"))
)
self.assertTrue(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "binary_file.txt"))
)

def test_auto_no_track_large_files_through_git_add(self):
def test_auto_no_track_files_through_git_add(self):
repo = Repository(WORKING_REPO_DIR)

# This content is 5MB (under 10MB)
Expand All @@ -1208,12 +1280,18 @@ def test_auto_no_track_large_files_through_git_add(self):
# This content is 20MB (over 10MB)
large_file = [100] * int(4e6)

# This content is binary (contains the null character)
binary_file = "\x00\x00\x00\x00"

with open(f"{WORKING_REPO_DIR}/large_file.txt", "w+") as f:
f.write(json.dumps(large_file))

with open(f"{WORKING_REPO_DIR}/small_file.txt", "w+") as f:
f.write(json.dumps(small_file))

with open(f"{WORKING_REPO_DIR}/binary_file.txt", "w+") as f:
f.write(binary_file)

repo.git_add(auto_lfs_track=False)

self.assertFalse(
Expand All @@ -1222,6 +1300,9 @@ def test_auto_no_track_large_files_through_git_add(self):
self.assertFalse(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "small_file.txt"))
)
self.assertFalse(
is_tracked_with_lfs(os.path.join(WORKING_REPO_DIR, "binary_file.txt"))
)

def test_auto_track_updates_removed_gitattributes(self):
repo = Repository(WORKING_REPO_DIR)
Expand Down