refactor: Extract code contribution stats logic to separate module

aider/commands.py

@@ -4,7 +4,7 @@ import re
 import subprocess
 import sys
 import tempfile
-from collections import OrderedDict
+from collections import OrderedDict, defaultdict
 from os.path import expanduser
 from pathlib import Path
 
@@ -23,6 +23,7 @@ from aider.repo import ANY_GIT_ERROR
 from aider.run_cmd import run_cmd
 from aider.scrape import Scraper, install_playwright
 from aider.utils import is_image_file
+from aider.stats import hash_len, get_all_commit_hashes_between_tags, get_commit_authors, get_counts_for_file
 
 from .dump import dump  # noqa: F401
 
@@ -1520,6 +1521,9 @@ class Commands:
                     self.io.tool_error("No main or master branch found. Please specify a revision.")
                     return
             source_revision, target_revision = args.split("..") if ".." in args else (args, "HEAD")
+            commits = get_all_commit_hashes_between_tags(source_revision, target_revision)
+            commits = [commit[:hash_len] for commit in commits] if commits else []
+            authors = get_commit_authors(commits)
 
             # Get files changed between revisions
             diff_files = self.coder.repo.repo.git.diff(
@@ -1535,49 +1539,29 @@ class Commands:
                 '.ttf', '.otf', '.woff', '.woff2', '.eot'  # fonts
             ))]
             self.io.tool_output(f"Found {len(files)} non-binary tracked files in the repository.")
-            
+
-            total_lines = 0
+            all_file_counts = {}
-            aider_lines = 0
+            grand_total = defaultdict(int)
-            
+            aider_total = 0
             for file in files:
-                try:
+                file_counts = get_counts_for_file(source_revision, target_revision, authors, file)
-                    # Run git blame for each file
+                if file_counts:
-                    blame_output = self.coder.repo.repo.git.blame(
+                    all_file_counts[file] = file_counts
-                        f"{source_revision}..{target_revision}", "-M", "-C", "--line-porcelain", "--", file
+                    for author, count in file_counts.items():
-                    )
+                        grand_total[author] += count
-                    
+                        if "(aider)" in author.lower():
-                    # Parse blame output
+                            aider_total += count
-                    for line in blame_output.split('filename'):
+            total_lines = sum(grand_total.values())
-                        total_lines += 1
+            aider_percentage = (aider_total / total_lines) * 100 if total_lines > 0 else 0
-                        for field in line.split('\n'):
-                            # Check author and committer lines for aider attribution
-                            author_match = False
-                            committer_match = False
-                            if field.startswith("author ") or field.startswith("committer "):
-                                author_match = "(aider)" in field.lower()
-                                committer_match = "(aider)" in field.lower()
-                            if author_match or committer_match:
-                                aider_lines += 1
-                    
-                except Exception as e:
-                    if "no such path" not in str(e).lower():
-                        self.io.tool_error(f"Error processing {file}: {e}")
 
             # Calculate percentages
             if total_lines > 0:
-                aider_percentage = (aider_lines / total_lines) * 100
+                # Output overall statistics
-                human_lines = total_lines - aider_lines
+                self.io.tool_output(f"\nAnalysis from {source_revision} to {target_revision}:")
-                human_percentage = (human_lines / total_lines) * 100
+                self.io.tool_output(f"Total lines analyzed: {total_lines:,}")
+                self.io.tool_output(f"Lines by aider: {aider_total:,} ({aider_percentage:.1f}%)")
+                self.io.tool_output(f"Lines by humans: {total_lines - aider_total:,} ({100 - aider_percentage:.1f}%)")
 
-                # Display results
-                self.io.tool_output("\nCode contribution statistics:")
-                self.io.tool_output(f"Total lines of code: {total_lines:,}")
-                self.io.tool_output(
-                    f"Human-written code: {human_lines:,} lines ({human_percentage:.1f}%)"
-                )
-                self.io.tool_output(
-                    f"Aider-written code: {aider_lines:,} lines ({aider_percentage:.1f}%)"
-                )
             else:
                 self.io.tool_output("No lines of code found in the repository.")
 

aider/stats.py

@@ -0,0 +1,84 @@
+import subprocess
+import sys
+
+from collections import defaultdict
+
+hash_len = len("44e6fefc2")
+
+def run(cmd):
+    # Get all commit hashes since the specified tag
+    result = subprocess.run(cmd, capture_output=True, text=True, check=True)
+    return result.stdout
+
+def get_all_commit_hashes_between_tags(start_tag, end_tag=None):
+    if end_tag:
+        res = run(["git", "rev-list", f"{start_tag}..{end_tag}"])
+    else:
+        res = run(["git", "rev-list", f"{start_tag}..HEAD"])
+
+    if res:
+        commit_hashes = res.strip().split("\n")
+        return commit_hashes
+
+def get_commit_authors(commits):
+    commit_to_author = dict()
+    for commit in commits:
+        author = run(["git", "show", "-s", "--format=%an", commit]).strip()
+        commit_message = run(["git", "show", "-s", "--format=%s", commit]).strip()
+        if commit_message.lower().startswith("aider:"):
+            author += " (aider)"
+        commit_to_author[commit] = author
+    return commit_to_author
+
+def get_counts_for_file(start_tag, end_tag, authors, fname):
+    try:
+        if end_tag:
+            text = run(
+                [
+                    "git",
+                    "blame",
+                    "-M100",  # Detect moved lines within a file with 100% similarity
+                    "-C100",  # Detect moves across files with 100% similarity
+                    "-C",  # Increase detection effort
+                    "-C",  # Increase detection effort even more
+                    "--abbrev=9",
+                    f"{start_tag}..{end_tag}",
+                    "--",
+                    fname,
+                ]
+            )
+        else:
+            text = run(
+                [
+                    "git",
+                    "blame",
+                    "-M100",  # Detect moved lines within a file with 100% similarity
+                    "-C100",  # Detect moves across files with 100% similarity
+                    "-C",  # Increase detection effort
+                    "-C",  # Increase detection effort even more
+                    "--abbrev=9",
+                    f"{start_tag}..HEAD",
+                    "--",
+                    fname,
+                ]
+            )
+        if not text:
+            return None
+        text = text.splitlines()
+        line_counts = defaultdict(int)
+        for line in text:
+            if line.startswith("^"):
+                continue
+            hsh = line[:hash_len]
+            author = authors.get(hsh, "Unknown")
+            line_counts[author] += 1
+
+        return dict(line_counts)
+    except subprocess.CalledProcessError as e:
+        if "no such path" in str(e).lower():
+            # File doesn't exist in this revision range, which is okay
+            return None
+        else:
+            # Some other error occurred
+            print(f"Warning: Unable to blame file {fname}. Error: {e}", file=sys.stderr)
+            return None

scripts/blame.py

@@ -2,8 +2,6 @@
 
 import argparse
 import os
-import subprocess
-import sys
 from collections import defaultdict
 from datetime import datetime
 from operator import itemgetter
@@ -12,6 +10,8 @@ import semver
 import yaml
 from tqdm import tqdm
 
+from aider.stats import get_all_commit_hashes_between_tags, run, hash_len, get_commit_authors, get_counts_for_file
+
 website_files = [
     "aider/website/index.html",
     "aider/website/share/index.md",
@@ -68,36 +68,6 @@ def blame(start_tag, end_tag=None):
     return all_file_counts, grand_total, total_lines, aider_total, aider_percentage, end_date
 
 
-def get_all_commit_hashes_between_tags(start_tag, end_tag=None):
-    if end_tag:
-        res = run(["git", "rev-list", f"{start_tag}..{end_tag}"])
-    else:
-        res = run(["git", "rev-list", f"{start_tag}..HEAD"])
-
-    if res:
-        commit_hashes = res.strip().split("\n")
-        return commit_hashes
-
-
-def run(cmd):
-    # Get all commit hashes since the specified tag
-    result = subprocess.run(cmd, capture_output=True, text=True, check=True)
-    return result.stdout
-
-
-def get_commit_authors(commits):
-    commit_to_author = dict()
-    for commit in commits:
-        author = run(["git", "show", "-s", "--format=%an", commit]).strip()
-        commit_message = run(["git", "show", "-s", "--format=%s", commit]).strip()
-        if commit_message.lower().startswith("aider:"):
-            author += " (aider)"
-        commit_to_author[commit] = author
-    return commit_to_author
-
-
-hash_len = len("44e6fefc2")
-
 
 def process_all_tags_since(start_tag):
     tags = get_all_tags_since(start_tag)