mirror of
https://github.com/newren/git-filter-repo.git
synced 2024-07-04 01:15:41 +02:00
filter-repo: split analysis reports into separate files
Signed-off-by: Elijah Newren <newren@gmail.com>
This commit is contained in:
Elijah Newren
parent
37c92d9352
commit
7048be2849
|
|
@ -1871,11 +1871,11 @@ def do_analysis(args, git_dir):
|
|||
results_tmp_dir = os.path.join(git_dir, 'filter-repo')
|
||||
if not os.path.isdir(results_tmp_dir):
|
||||
os.mkdir(results_tmp_dir)
|
||||
reportfile = os.path.join(results_tmp_dir,
|
||||
"repo-analysis-{}.txt".format(time.strftime("%F")))
|
||||
if not args.force and os.path.isfile(reportfile):
|
||||
reportdir = os.path.join(results_tmp_dir, "analysis")
|
||||
if not args.force and os.path.isdir(reportdir):
|
||||
raise SystemExit("Error: {} already exists; refusing to overwrite!".
|
||||
format(reportfile))
|
||||
format(reportdir))
|
||||
os.mkdir(reportdir)
|
||||
|
||||
# Now gather the data we need
|
||||
gather_data(args)
|
||||
|
|
@ -1931,25 +1931,8 @@ def do_analysis(args, git_dir):
|
|||
else:
|
||||
dir_deleted_data[dirname] = when
|
||||
|
||||
with open(reportfile, 'w') as f:
|
||||
with open(os.path.join(reportdir, "README"), 'w') as f:
|
||||
# Give a basic overview of this file
|
||||
f.write("== Table of Contents ==\n")
|
||||
f.write(" * Overal Statistics\n")
|
||||
f.write(" * Caveats\n")
|
||||
f.write(" * File renames\n")
|
||||
f.write(" * Directory sizes\n")
|
||||
f.write(" * Deleted directories\n")
|
||||
f.write(" * All directories\n")
|
||||
f.write(" * Filename extension sizes\n")
|
||||
f.write(" * Deleted extensions\n")
|
||||
f.write(" * All extensions\n")
|
||||
f.write(" * Path sizes (accumulated across commits)\n")
|
||||
f.write(" * Deleted paths\n")
|
||||
f.write(" * All paths\n")
|
||||
f.write(" * Files by sha and associated pathnames\n")
|
||||
f.write("\n")
|
||||
|
||||
# Provide total unpacked size
|
||||
f.write("== Overal Statistics ==\n")
|
||||
f.write(" Number of commits: {}\n".format(args.num_commits))
|
||||
f.write(" Number of filenames: {}\n".format(len(path_size)))
|
||||
|
|
@ -1983,6 +1966,7 @@ def do_analysis(args, git_dir):
|
|||
This makes it dependent on topological ordering, but generally gives
|
||||
the "right" answer.
|
||||
"""[1:]))
|
||||
f.write("\n")
|
||||
f.write("=== Renames ===\n")
|
||||
f.write(textwrap.dedent("""
|
||||
Renames share the same non-binary nature that deletions do, plus
|
||||
|
|
@ -2013,23 +1997,22 @@ def do_analysis(args, git_dir):
|
|||
"""[1:]))
|
||||
f.write("\n")
|
||||
|
||||
# Equivalence classes for names, so if folks only want to keep a
|
||||
# certain set of paths, they know the old names they want to include
|
||||
# too.
|
||||
f.write("== File renames ==\n")
|
||||
# Equivalence classes for names, so if folks only want to keep a
|
||||
# certain set of paths, they know the old names they want to include
|
||||
# too.
|
||||
with open(os.path.join(reportdir, "renames.txt"), 'w') as f:
|
||||
seen = set()
|
||||
for pathname,equiv_group in sorted(args.stats['equivalence'].iteritems(),
|
||||
key=lambda x:x[1]):
|
||||
if equiv_group in seen:
|
||||
continue
|
||||
seen.add(equiv_group)
|
||||
f.write(" {} ->\n ".format(equiv_group[0]) +
|
||||
"\n ".join(equiv_group[1:]) +
|
||||
f.write("{} ->\n ".format(equiv_group[0]) +
|
||||
"\n ".join(equiv_group[1:]) +
|
||||
"\n")
|
||||
f.write("\n")
|
||||
|
||||
# List directories in reverse sorted order of unpacked size
|
||||
f.write("== Directory sizes ==\n")
|
||||
# List directories in reverse sorted order of unpacked size
|
||||
with open(os.path.join(reportdir, "directories-deleted-sizes.txt"), 'w') as f:
|
||||
f.write("=== Deleted directories by reverse size ===\n")
|
||||
f.write("Format: size (bytes), date deleted, directory name\n")
|
||||
for dirname, size in sorted(dir_size.iteritems(),
|
||||
|
|
@ -2038,17 +2021,17 @@ def do_analysis(args, git_dir):
|
|||
f.write(" {:10d} {:10s} {}\n".format(size,
|
||||
datestr(dir_deleted_data[dirname]),
|
||||
dirname or '<toplevel>'))
|
||||
f.write("\n")
|
||||
|
||||
with open(os.path.join(reportdir, "directories-all-sizes.txt"), 'w') as f:
|
||||
f.write("=== All directories by reverse size ===\n")
|
||||
f.write("Format: size (bytes), date deleted, directory name\n")
|
||||
for dirname, size in sorted(dir_size.iteritems(),
|
||||
key=lambda x:x[1], reverse=True):
|
||||
f.write(" {:10d} {:10s} {}\n".format(size, datestr(dir_deleted_data[dirname]),
|
||||
dirname or '<toplevel>'))
|
||||
f.write("\n")
|
||||
|
||||
# List extensions in reverse sorted order of unpacked size
|
||||
f.write("== Filename extension sizes ==\n")
|
||||
# List extensions in reverse sorted order of unpacked size
|
||||
with open(os.path.join(reportdir, "extensions-deleted-sizes.txt"), 'w') as f:
|
||||
f.write("=== Deleted extensions by reverse size ===\n")
|
||||
f.write("Format: size (bytes), date deleted, extension name\n")
|
||||
for extname, size in sorted(ext_size.iteritems(),
|
||||
|
|
@ -2057,7 +2040,8 @@ def do_analysis(args, git_dir):
|
|||
f.write(" {:10d} {:10s} {}\n".format(size,
|
||||
datestr(ext_deleted_data[extname]),
|
||||
extname or '<no extension>'))
|
||||
f.write("\n")
|
||||
|
||||
with open(os.path.join(reportdir, "extensions-all-sizes.txt"), 'w') as f:
|
||||
f.write("=== All extensions by reverse size ===\n")
|
||||
f.write("Format: size (bytes), date deleted, extension name\n")
|
||||
for extname, size in sorted(ext_size.iteritems(),
|
||||
|
|
@ -2065,27 +2049,27 @@ def do_analysis(args, git_dir):
|
|||
f.write(" {:10d} {:10s} {}\n".format(size,
|
||||
datestr(ext_deleted_data[extname]),
|
||||
extname or '<no extension>'))
|
||||
f.write("\n")
|
||||
|
||||
# List files in reverse sorted order of unpacked size
|
||||
f.write("== Path sizes (accumulated across commits) ==\n")
|
||||
f.write("=== Deleted paths by reverse size ===\n")
|
||||
# List files in reverse sorted order of unpacked size
|
||||
with open(os.path.join(reportdir, "path-deleted-sizes.txt"), 'w') as f:
|
||||
f.write("=== Deleted paths by reverse accumulated size ===\n")
|
||||
f.write("Format: size (bytes), date deleted, path name(s)\n")
|
||||
for pathname, size in sorted(path_size.iteritems(),
|
||||
key=lambda x:x[1], reverse=True):
|
||||
when = args.stats['deletions'].get(pathname, None)
|
||||
if when:
|
||||
f.write(" {:10d} {:10s} {}\n".format(size, datestr(when), pathname))
|
||||
f.write("\n")
|
||||
f.write("=== All paths by reverse size ===\n")
|
||||
|
||||
with open(os.path.join(reportdir, "path-all-sizes.txt"), 'w') as f:
|
||||
f.write("=== All paths by reverse accumulated size ===\n")
|
||||
f.write("Format: size (bytes), date deleted, pathectory name\n")
|
||||
for pathname, size in sorted(path_size.iteritems(),
|
||||
key=lambda x:x[1], reverse=True):
|
||||
when = args.stats['deletions'].get(pathname, None)
|
||||
f.write(" {:10d} {:10s} {}\n".format(size, datestr(when), pathname))
|
||||
f.write("\n")
|
||||
|
||||
# List of filenames and sizes in descending order
|
||||
# List of filenames and sizes in descending order
|
||||
with open(os.path.join(reportdir, "blob-shas-and-paths.txt"), 'w') as f:
|
||||
f.write("== Files by sha and associated pathnames in reverse size ==\n")
|
||||
f.write("Format: sha, size (bytes), filename(s) object stored as\n")
|
||||
for sha, size in sorted(args.size.iteritems(), key=lambda x:x[1],
|
||||
|
|
@ -2100,8 +2084,9 @@ def do_analysis(args, git_dir):
|
|||
else:
|
||||
names_with_sha = sorted(list(names_with_sha))
|
||||
f.write(" {} {:9d} {}\n".format(sha, size, names_with_sha))
|
||||
f.write("\n")
|
||||
print("Report written to {}".format(reportfile))
|
||||
|
||||
# Notify the user where they can find the reports
|
||||
print("Reports written to {}".format(reportdir))
|
||||
|
||||
def tweak_commit(args, commit):
|
||||
def filename_matches(path_expression, pathname):
|
||||
|
|
|
|||
Loading…
Reference in New Issue
Block a user