dvc-utils 0.0.3__tar.gz → 0.0.5__tar.gz

{dvc-utils-0.0.3 → dvc-utils-0.0.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: dvc-utils
-Version: 0.0.3
+Version: 0.0.5
 Summary: CLI for diffing DVC files at two commits (or one commit vs. current worktree), optionally passing both through another command first
 Home-page: https://github.com/runsascoded/dvc-utils
 Author: Ryan Williams
@@ -29,6 +29,7 @@ pip install dvc-utils
 ```
 
 ## Usage <a id="usage"></a>
+<!-- `bmdf -- dvc-utils --help` -->
 ```bash
 dvc-utils --help
 # Usage: dvc-utils [OPTIONS] COMMAND [ARGS]...
@@ -42,6 +43,7 @@ dvc-utils --help
 ```
 
 ### `dvc-utils diff` <a id="dvc-utils-diff"></a>
+<!-- `bmdf -- dvc-utils diff --help` -->
 ```bash
 dvc-utils diff --help
 # Usage: dvc-utils diff [OPTIONS] [cmd...] <path>
@@ -59,11 +61,15 @@ dvc-utils diff --help
 #   optional) at HEAD (last committed value) vs. the current worktree content.
 #
 # Options:
-#   -r, --refspec TEXT  <commit 1>..<commit 2> (compare two commits) or <commit>
-#                       (compare <commit> to the worktree)
-#   -S, --no-shell      Don't pass `shell=True` to Python `subprocess`es
-#   -v, --verbose       Log intermediate commands to stderr
-#   --help              Show this message and exit.
+#   -c, --color              Colorize the output
+#   -r, --refspec TEXT       <commit 1>..<commit 2> (compare two commits) or
+#                            <commit> (compare <commit> to the worktree)
+#   -S, --no-shell           Don't pass `shell=True` to Python `subprocess`es
+#   -U, --unified INTEGER    Number of lines of context to show (passes through
+#                            to `diff`)
+#   -v, --verbose            Log intermediate commands to stderr
+#   -w, --ignore-whitespace  Ignore whitespace differences (pass `-w` to `diff`)
+#   --help                   Show this message and exit.
 ```
 
 ## Examples <a id="examples"></a>

{dvc-utils-0.0.3 → dvc-utils-0.0.5}/README.md

@@ -18,6 +18,7 @@ pip install dvc-utils
 ```
 
 ## Usage <a id="usage"></a>
+<!-- `bmdf -- dvc-utils --help` -->
 ```bash
 dvc-utils --help
 # Usage: dvc-utils [OPTIONS] COMMAND [ARGS]...
@@ -31,6 +32,7 @@ dvc-utils --help
 ```
 
 ### `dvc-utils diff` <a id="dvc-utils-diff"></a>
+<!-- `bmdf -- dvc-utils diff --help` -->
 ```bash
 dvc-utils diff --help
 # Usage: dvc-utils diff [OPTIONS] [cmd...] <path>
@@ -48,11 +50,15 @@ dvc-utils diff --help
 #   optional) at HEAD (last committed value) vs. the current worktree content.
 #
 # Options:
-#   -r, --refspec TEXT  <commit 1>..<commit 2> (compare two commits) or <commit>
-#                       (compare <commit> to the worktree)
-#   -S, --no-shell      Don't pass `shell=True` to Python `subprocess`es
-#   -v, --verbose       Log intermediate commands to stderr
-#   --help              Show this message and exit.
+#   -c, --color              Colorize the output
+#   -r, --refspec TEXT       <commit 1>..<commit 2> (compare two commits) or
+#                            <commit> (compare <commit> to the worktree)
+#   -S, --no-shell           Don't pass `shell=True` to Python `subprocess`es
+#   -U, --unified INTEGER    Number of lines of context to show (passes through
+#                            to `diff`)
+#   -v, --verbose            Log intermediate commands to stderr
+#   -w, --ignore-whitespace  Ignore whitespace differences (pass `-w` to `diff`)
+#   --help                   Show this message and exit.
 ```
 
 ## Examples <a id="examples"></a>

dvc-utils-0.0.5/dvc_utils/main.py

@@ -0,0 +1,233 @@
+from functools import cache
+from os import environ as env, getcwd
+from os.path import join, relpath
+import shlex
+from subprocess import Popen, PIPE
+from typing import Optional, Tuple
+
+from click import option, argument, group
+import click
+import yaml
+from utz import process, singleton, err
+
+from dvc_utils.named_pipes import named_pipes
+
+
+@group()
+def cli():
+    pass
+
+
+def dvc_paths(path: str) -> Tuple[str, str]:
+    if path.endswith('.dvc'):
+        dvc_path = path
+        path = dvc_path[:-len('.dvc')]
+    else:
+        dvc_path = f'{path}.dvc'
+    return path, dvc_path
+
+
+@cache
+def get_git_root() -> str:
+    return process.line('git', 'rev-parse', '--show-toplevel', log=False)
+
+
+@cache
+def get_dir_path() -> str:
+    return relpath(getcwd(), get_git_root())
+
+
+@cache
+def dvc_cache_dir(log: bool = False) -> str:
+    dvc_cache_relpath = env.get('DVC_UTILS_CACHE_DIR')
+    if dvc_cache_relpath:
+        return join(get_git_root(), dvc_cache_relpath)
+    else:
+        return process.line('dvc', 'cache', 'dir', log=log)
+
+
+def dvc_md5(git_ref: str, dvc_path: str, log: bool = False) -> str:
+    dir_path = get_dir_path()
+    dir_path = '' if dir_path == '.' else f'{dir_path}/'
+    dvc_spec = process.output('git', 'show', f'{git_ref}:{dir_path}{dvc_path}', log=err if log else None)
+    dvc_obj = yaml.safe_load(dvc_spec)
+    out = singleton(dvc_obj['outs'], dedupe=False)
+    md5 = out['md5']
+    return md5
+
+
+def dvc_cache_path(ref: str, dvc_path: Optional[str] = None, log: bool = False) -> str:
+    if dvc_path:
+        md5 = dvc_md5(ref, dvc_path, log=log)
+    elif ':' in ref:
+        git_ref, dvc_path = ref.split(':', 1)
+        md5 = dvc_md5(git_ref, dvc_path, log=log)
+    else:
+        md5 = ref
+    dirname = md5[:2]
+    basename = md5[2:]
+    return join(dvc_cache_dir(log=log), 'files', 'md5', dirname, basename)
+
+
+def diff_cmds(
+    cmds1: list[str],
+    cmds2: list[str],
+    verbose: bool = False,
+    color: bool = False,
+    unified: int | None = None,
+    ignore_whitespace: bool = False,
+    **kwargs,
+):
+    """Run two sequences of piped commands and diff their output.
+
+    Args:
+        cmds1: First sequence of commands to pipe together
+        cmds2: Second sequence of commands to pipe together
+        verbose: Whether to print commands being executed
+        color: Whether to show colored diff output
+        unified: Number of unified context lines, or None
+        ignore_whitespace: Whether to ignore whitespace changes
+        **kwargs: Additional arguments passed to subprocess.Popen
+
+    Each command sequence will be piped together before being compared.
+    For example, if cmds1 = ['cat foo.txt', 'sort'], the function will
+    execute 'cat foo.txt | sort' before comparing with cmds2's output.
+
+    Adapted from https://stackoverflow.com/a/28840955"""
+    with named_pipes(n=2) as pipes:
+        (pipe1, pipe2) = pipes
+        diff_cmd = [
+            'diff',
+            *(['-w'] if ignore_whitespace else []),
+            *(['-U', str(unified)] if unified is not None else []),
+            *(['--color=always'] if color else []),
+            pipe1,
+            pipe2,
+        ]
+        diff = Popen(diff_cmd)
+        processes = []
+
+        for pipe, cmds in ((pipe1, cmds1), (pipe2, cmds2)):
+            if verbose:
+                err(f"Running pipeline: {' | '.join(cmds)}")
+
+            # Create the pipeline of processes
+            prev_process = None
+            for i, cmd in enumerate(cmds):
+                is_last = i + 1 == len(cmds)
+
+                # For the first process, take input from the original source
+                stdin = None if prev_process is None else prev_process.stdout
+
+                # For the last process, output to the named pipe
+                if is_last:
+                    with open(pipe, 'wb', 0) as pipe_fd:
+                        proc = Popen(
+                            cmd,
+                            stdin=stdin,
+                            stdout=pipe_fd,
+                            close_fds=True,
+                            **kwargs
+                        )
+                # For intermediate processes, output to a pipe
+                else:
+                    proc = Popen(
+                        cmd,
+                        stdin=stdin,
+                        stdout=PIPE,
+                        close_fds=True,
+                        **kwargs
+                    )
+
+                if prev_process is not None:
+                    prev_process.stdout.close()
+
+                processes.append(proc)
+                prev_process = proc
+
+        for p in [diff] + processes:
+            p.wait()
+
+
+@cli.command('diff', short_help='Diff a DVC-tracked file at two commits (or one commit vs. current worktree), optionally passing both through another command first')
+@option('-c', '--color', is_flag=True, help='Colorize the output')
+@option('-r', '--refspec', default='HEAD', help='<commit 1>..<commit 2> (compare two commits) or <commit> (compare <commit> to the worktree)')
+@option('-S', '--no-shell', is_flag=True, help="Don't pass `shell=True` to Python `subprocess`es")
+@option('-U', '--unified', type=int, help='Number of lines of context to show (passes through to `diff`)')
+@option('-v', '--verbose', is_flag=True, help="Log intermediate commands to stderr")
+@option('-w', '--ignore-whitespace', is_flag=True, help="Ignore whitespace differences (pass `-w` to `diff`)")
+@option('-x', '--exec-cmd', 'exec_cmds', multiple=True, help='Command(s) to execute before diffing; alternate syntax to passing commands as positional arguments')
+@argument('args', metavar='[exec_cmd...] <path>', nargs=-1)
+def dvc_utils_diff(
+    color: bool,
+    refspec: str | None,
+    no_shell: bool,
+    unified: int | None,
+    verbose: bool,
+    ignore_whitespace: bool,
+    exec_cmds: Tuple[str, ...],
+    args: Tuple[str, ...],
+):
+    """Diff a file at two commits (or one commit vs. current worktree), optionally passing both through `cmd` first
+
+    Examples:
+
+    dvc-utils diff -r HEAD^..HEAD wc -l foo.dvc  # Compare the number of lines (`wc -l`) in `foo` (the file referenced by `foo.dvc`) at the previous vs. current commit (`HEAD^..HEAD`).
+
+    dvc-utils diff md5sum foo  # Diff the `md5sum` of `foo` (".dvc" extension is optional) at HEAD (last committed value) vs. the current worktree content.
+    """
+    if not args:
+        raise click.UsageError('Must specify [cmd...] <path>')
+
+    shell = not no_shell
+    *cmds, path = args
+    cmds = list(exec_cmds) + cmds
+
+    path, dvc_path = dvc_paths(path)
+
+    pcs = refspec.split('..', 1)
+    if len(pcs) == 1:
+        before = pcs[0]
+        after = None
+    elif len(pcs) == 2:
+        before, after = pcs
+    else:
+        raise ValueError(f"Invalid refspec: {refspec}")
+
+    log = err if verbose else False
+    before_path = dvc_cache_path(before, dvc_path, log=log)
+    after_path = path if after is None else dvc_cache_path(after, dvc_path, log=log)
+
+    if cmds:
+        cmd, *sub_cmds = cmds
+        if not shell:
+            sub_cmds = [ shlex.split(c) for c in sub_cmds ]
+            before_cmds = [
+                shlex.split(f'{cmd} {before_path}'),
+                *sub_cmds,
+            ]
+            after_cmds = [
+                shlex.split(f'{cmd} {after_path}'),
+                *sub_cmds,
+            ]
+            shell_kwargs = {}
+        else:
+            before_cmds = [ f'{cmd} {before_path}', *sub_cmds ]
+            after_cmds = [ f'{cmd} {after_path}', *sub_cmds ]
+            shell_kwargs = dict(shell=shell)
+
+        diff_cmds(
+            before_cmds,
+            after_cmds,
+            verbose=verbose,
+            color=color,
+            unified=unified,
+            ignore_whitespace=ignore_whitespace,
+            **shell_kwargs,
+        )
+    else:
+        process.run('diff', before_path, after_path, log=log)
+
+
+if __name__ == '__main__':
+    cli()

{dvc-utils-0.0.3 → dvc-utils-0.0.5}/dvc_utils.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: dvc-utils
-Version: 0.0.3
+Version: 0.0.5
 Summary: CLI for diffing DVC files at two commits (or one commit vs. current worktree), optionally passing both through another command first
 Home-page: https://github.com/runsascoded/dvc-utils
 Author: Ryan Williams
@@ -29,6 +29,7 @@ pip install dvc-utils
 ```
 
 ## Usage <a id="usage"></a>
+<!-- `bmdf -- dvc-utils --help` -->
 ```bash
 dvc-utils --help
 # Usage: dvc-utils [OPTIONS] COMMAND [ARGS]...
@@ -42,6 +43,7 @@ dvc-utils --help
 ```
 
 ### `dvc-utils diff` <a id="dvc-utils-diff"></a>
+<!-- `bmdf -- dvc-utils diff --help` -->
 ```bash
 dvc-utils diff --help
 # Usage: dvc-utils diff [OPTIONS] [cmd...] <path>
@@ -59,11 +61,15 @@ dvc-utils diff --help
 #   optional) at HEAD (last committed value) vs. the current worktree content.
 #
 # Options:
-#   -r, --refspec TEXT  <commit 1>..<commit 2> (compare two commits) or <commit>
-#                       (compare <commit> to the worktree)
-#   -S, --no-shell      Don't pass `shell=True` to Python `subprocess`es
-#   -v, --verbose       Log intermediate commands to stderr
-#   --help              Show this message and exit.
+#   -c, --color              Colorize the output
+#   -r, --refspec TEXT       <commit 1>..<commit 2> (compare two commits) or
+#                            <commit> (compare <commit> to the worktree)
+#   -S, --no-shell           Don't pass `shell=True` to Python `subprocess`es
+#   -U, --unified INTEGER    Number of lines of context to show (passes through
+#                            to `diff`)
+#   -v, --verbose            Log intermediate commands to stderr
+#   -w, --ignore-whitespace  Ignore whitespace differences (pass `-w` to `diff`)
+#   --help                   Show this message and exit.
 ```
 
 ## Examples <a id="examples"></a>

{dvc-utils-0.0.3 → dvc-utils-0.0.5}/setup.py

@@ -2,7 +2,7 @@ from setuptools import setup
 
 setup(
     name='dvc-utils',
-    version="0.0.3",
+    version="0.0.5",
     description="CLI for diffing DVC files at two commits (or one commit vs. current worktree), optionally passing both through another command first",
     long_description=open("README.md").read(),
     long_description_content_type="text/markdown",

dvc-utils-0.0.3/dvc_utils/main.py

@@ -1,148 +0,0 @@
-from functools import cache
-from os import environ as env, getcwd
-
-from typing import Optional, Tuple
-
-import shlex
-from os.path import join, relpath
-
-from click import option, argument, group
-from subprocess import Popen
-
-import click
-import yaml
-from utz import process, singleton, err
-
-from dvc_utils.named_pipes import named_pipes
-
-
-@group()
-def cli():
-    pass
-
-
-def dvc_paths(path: str) -> Tuple[str, str]:
-    if path.endswith('.dvc'):
-        dvc_path = path
-        path = dvc_path[:-len('.dvc')]
-    else:
-        dvc_path = f'{path}.dvc'
-    return path, dvc_path
-
-
-@cache
-def get_git_root() -> str:
-    return process.line('git', 'rev-parse', '--show-toplevel', log=False)
-
-
-@cache
-def get_dir_path() -> str:
-    return relpath(getcwd(), get_git_root())
-
-
-@cache
-def dvc_cache_dir(log: bool = False) -> str:
-    dvc_cache_relpath = env.get('DVC_UTILS_CACHE_DIR')
-    if dvc_cache_relpath:
-        return join(get_git_root(), dvc_cache_relpath)
-    else:
-        return process.line('dvc', 'cache', 'dir', log=log)
-
-
-def dvc_md5(git_ref: str, dvc_path: str, log: bool = False) -> str:
-    dir_path = get_dir_path()
-    dir_path = '' if dir_path == '.' else f'{dir_path}/'
-    dvc_spec = process.output('git', 'show', f'{git_ref}:{dir_path}{dvc_path}', log=log)
-    dvc_obj = yaml.safe_load(dvc_spec)
-    out = singleton(dvc_obj['outs'], dedupe=False)
-    md5 = out['md5']
-    return md5
-
-
-def dvc_cache_path(ref: str, dvc_path: Optional[str] = None, log: bool = False) -> str:
-    if dvc_path:
-        md5 = dvc_md5(ref, dvc_path, log=log)
-    elif ':' in ref:
-        git_ref, dvc_path = ref.split(':', 1)
-        md5 = dvc_md5(git_ref, dvc_path, log=log)
-    else:
-        md5 = ref
-    dirname = md5[:2]
-    basename = md5[2:]
-    return join(dvc_cache_dir(log=log), 'files', 'md5', dirname, basename)
-
-
-def diff_cmds(cmd1: str, cmd2: str, verbose: bool = False, **kwargs):
-    """Run two commands and diff their output.
-
-    Adapted from https://stackoverflow.com/a/28840955"""
-    with named_pipes(n=2) as pipes:
-        (pipe1, pipe2) = pipes
-        diff = Popen(['diff'] + pipes)
-        processes = []
-        for path, cmd in ((pipe1, cmd1), (pipe2, cmd2)):
-            with open(path, 'wb', 0) as pipe:
-                if verbose:
-                    err(f"Running: {cmd}")
-                processes.append(Popen(cmd, stdout=pipe, close_fds=True, **kwargs))
-        for p in [diff] + processes:
-            p.wait()
-
-
-@cli.command('diff', short_help='Diff a DVC-tracked file at two commits (or one commit vs. current worktree), optionally passing both through another command first')
-@option('-r', '--refspec', default='HEAD', help='<commit 1>..<commit 2> (compare two commits) or <commit> (compare <commit> to the worktree)')
-@option('-S', '--no-shell', is_flag=True, help="Don't pass `shell=True` to Python `subprocess`es")
-@option('-v', '--verbose', is_flag=True, help="Log intermediate commands to stderr")
-@argument('args', metavar='[cmd...] <path>', nargs=-1)
-def dvc_utils_diff(refspec, no_shell, verbose, args):
-    """Diff a file at two commits (or one commit vs. current worktree), optionally passing both through `cmd` first
-
-    Examples:
-
-    dvc-utils diff -r HEAD^..HEAD wc -l foo.dvc  # Compare the number of lines (`wc -l`) in `foo` (the file referenced by `foo.dvc`) at the previous vs. current commit (`HEAD^..HEAD`).
-
-    dvc-utils diff md5sum foo  # Diff the `md5sum` of `foo` (".dvc" extension is optional) at HEAD (last committed value) vs. the current worktree content.
-    """
-    if not args:
-        raise click.UsageError('Must specify [cmd...] <path>')
-
-    shell = not no_shell
-    if len(args) == 2:
-        cmd, path = args
-        cmd = shlex.split(cmd)
-    elif len(args) == 1:
-        cmd = None
-        path, = args
-    else:
-        raise click.UsageError('Maximum 2 positional args: [cmd] <path>')
-
-    path, dvc_path = dvc_paths(path)
-
-    pcs = refspec.split('..', 1)
-    if len(pcs) == 1:
-        before = pcs[0]
-        after = None
-    elif len(pcs) == 2:
-        before, after = pcs
-    else:
-        raise ValueError(f"Invalid refspec: {refspec}")
-
-    log = err if verbose else False
-    before_path = dvc_cache_path(before, dvc_path, log=log)
-    after_path = path if after is None else dvc_cache_path(after, dvc_path, log=log)
-
-    if cmd:
-        def args(path: str):
-            arr = cmd + [path]
-            return shlex.join(arr) if shell else arr
-
-        shell_kwargs = dict(shell=shell) if shell else {}
-        before_cmd = args(before_path)
-        after_cmd = args(after_path)
-        diff_cmds(before_cmd, after_cmd, verbose=verbose, **shell_kwargs)
-    else:
-        process.run('diff', before_path, after_path, log=log)
-
-
-if __name__ == '__main__':
-    cli()