ci: add automated benchmark data push after merge

.github/scripts/benchmark_formatter.py

@@ -9,6 +9,11 @@
     lines = p.read_text(encoding="utf-8").splitlines()
     processed_lines = []
     in_code = False
+    delta_col = None  # record "Diff" column start per table
+    align_hint = None  # derived from benchstat header last pipe position
+
+    ALIGN_COLUMN = 60  # fallback alignment when header not found
+
     def strip_worker_suffix(text: str) -> str:
         return re.sub(r'(\S+?)-\d+(\s|$)', r'\1\2', text)
 
@@ -117,8 +122,6 @@ def extract_two_numbers(tokens):
     # Diff column width ~12 chars (e.g. "+100.00% 🚀")
     right_boundary = diff_col_start + 14
 
-    # Reset code fence tracking state for Pass 1
-    in_code = False
     for line in lines:
 
         if line.strip() == "```":
@@ -146,6 +149,12 @@ def extract_two_numbers(tokens):
             stripped_header = re.sub(r'\s+Delta\b', '', stripped_header, flags=re.IGNORECASE)
 
             # Pad to diff_col_start
+            padding = diff_col_start - len(stripped_header)
+            if padding < 2: 
+                padding = 2 # minimum spacing
+                # If header is wider than data (unlikely but possible), adjust diff_col_start
+                # But for now let's trust max_content_width or just append
+
             if len(stripped_header) < diff_col_start:
                 new_header = stripped_header + " " * (diff_col_start - len(stripped_header))
             else:
@@ -169,6 +178,7 @@ def extract_two_numbers(tokens):
             processed_lines.append(line)
             continue
 
+        original_line = line
         line = strip_worker_suffix(line)
         tokens = line.split()
         if not tokens:

.github/scripts/format_benchmark_data.py

@@ -0,0 +1,101 @@
+import json
+import os
+import sys
+import datetime
+import re
+
+
+def normalize_name(name):
+    # Remove "Benchmark" prefix
+    name = re.sub(r"^Benchmark", "", name)
+    # Remove -N suffix (GOMAXPROCS)
+    name = re.sub(r"-\d+$", "", name)
+
+    parts = re.split(r'[/_]', name)
+    new_parts = []
+    for p in parts:
+        if p.lower() in ["rbac", "abac", "acl", "api", "rest"]:
+            new_parts.append(p.upper())
+        else:
+            new_parts.append(p.capitalize())
+    return "".join(new_parts)
+
+
+def main():
+    if len(sys.argv) < 3:
+        print("Usage: python format_benchmark_data.py input.txt output.json")
+        sys.exit(1)
+
+    input_path = sys.argv[1]
+    output_path = sys.argv[2]
+
+    # Get commit info from environment variables
+    # These should be set in the GitHub Action
+    commit_info = {
+        "author": {
+            "email": os.environ.get("COMMIT_AUTHOR_EMAIL", ""),
+            "name": os.environ.get("COMMIT_AUTHOR_NAME", ""),
+            "username": os.environ.get("COMMIT_AUTHOR_USERNAME", ""),
+        },
+        "committer": {
+            "email": os.environ.get("COMMIT_COMMITTER_EMAIL", ""),
+            "name": os.environ.get("COMMIT_COMMITTER_NAME", ""),
+            "username": os.environ.get("COMMIT_COMMITTER_USERNAME", ""),
+        },
+        "distinct": True,  # Assuming true for push to master
+        "id": os.environ.get("COMMIT_ID", ""),
+        "message": os.environ.get("COMMIT_MESSAGE", ""),
+        "timestamp": os.environ.get("COMMIT_TIMESTAMP", ""),
+        "tree_id": os.environ.get("COMMIT_TREE_ID", ""),
+        "url": os.environ.get("COMMIT_URL", ""),
+    }
+
+    # Get CPU count
+    cpu_count = os.cpu_count() or 1
+
+    benches = []
+
+    try:
+        with open(input_path, "r", encoding="utf-8") as f:
+            lines = f.readlines()
+
+        for line in lines:
+            # Parse Go benchmark output: BenchmarkName-8 10000 123 ns/op
+            # Also handle lines with 4 columns if any, but standard is: Name Iterations Value Unit
+            match = re.search(r'^(Benchmark\S+)\s+(\d+)\s+([\d\.]+)\s+ns/op', line)
+            if match:
+                name = match.group(1)
+                iterations = int(match.group(2))
+                val_ns = float(match.group(3))
+
+                # Format extra info
+                extra = f"{iterations} times"
+
+                # Create entry
+                benches.append({
+                    "name": normalize_name(name),
+                    "value": round(val_ns, 2),
+                    "unit": "ns/op",
+                    "extra": extra
+                })
+
+    except Exception as e:
+        print(f"Error processing {input_path}: {e}")
+        sys.exit(1)
+
+    output_data = {
+        "commit": commit_info,
+        "date": int(datetime.datetime.now().timestamp() * 1000),  # Current timestamp in ms
+        "tool": "go",
+        "procs": cpu_count,
+        "benches": benches,
+    }
+
+    with open(output_path, "w", encoding="utf-8") as f:
+        json.dump(output_data, f, indent=2)
+
+    print(f"Successfully formatted benchmark data to {output_path}")
+
+
+if __name__ == "__main__":
+    main()

.github/scripts/merge_benchmarks.py

@@ -0,0 +1,36 @@
+import json
+import sys
+import glob
+
+
+def merge_jsons(output_file, input_files):
+    merged_data = None
+    all_benchmarks = []
+
+    for file_path in input_files:
+        with open(file_path, "r", encoding="utf-8") as f:
+            data = json.load(f)
+            if merged_data is None:
+                merged_data = data
+            all_benchmarks.extend(data.get("benchmarks", []))
+
+    if merged_data:
+        merged_data["benchmarks"] = all_benchmarks
+        with open(output_file, "w", encoding="utf-8") as f:
+            json.dump(merged_data, f, indent=4)
+
+
+if __name__ == "__main__":
+    if len(sys.argv) < 3:
+        print("Usage: python merge_benchmarks.py output.json input1.json input2.json ...")
+        sys.exit(1)
+
+    output_file = sys.argv[1]
+    input_files = sys.argv[2:]
+
+    # Expand globs if shell didn't
+    expanded_inputs = []
+    for p in input_files:
+        expanded_inputs.extend(glob.glob(p))
+
+    merge_jsons(output_file, expanded_inputs)

.github/scripts/update_data_js.py

@@ -0,0 +1,82 @@
+import json
+import sys
+import os
+import re
+import time
+
+
+def load_data_js(filepath):
+    if not os.path.exists(filepath):
+        return {"lastUpdate": 0, "repoUrl": "https://github.com/casbin/casbin", "entries": {}}
+
+    with open(filepath, "r", encoding="utf-8") as f:
+        content = f.read()
+
+    # Strip window.BENCHMARK_DATA =
+    match = re.search(r"window\.BENCHMARK_DATA\s*=\s*({.*});?", content, re.DOTALL)
+    if match:
+        try:
+            return json.loads(match.group(1))
+        except json.JSONDecodeError:
+            print("Error decoding JSON from data.js", file=sys.stderr)
+            sys.exit(1)
+    return {"lastUpdate": 0, "repoUrl": "https://github.com/casbin/casbin", "entries": {}}
+
+
+def save_data_js(filepath, data):
+    content = f"window.BENCHMARK_DATA = {json.dumps(data, indent=4)};"
+    with open(filepath, "w", encoding="utf-8") as f:
+        f.write(content)
+
+
+def main():
+    if len(sys.argv) < 3:
+        print("Usage: python update_data_js.py benchmark_result.json data.js")
+        sys.exit(1)
+
+    bench_file = sys.argv[1]
+    data_js_file = sys.argv[2]
+
+    with open(bench_file, "r", encoding="utf-8") as f:
+        bench_data = json.load(f)
+
+    js_data = load_data_js(data_js_file)
+
+    # Construct new entry
+    commit_info = bench_data.get("commit_info", {})
+
+    entry = {
+        "commit": {
+            "time": commit_info.get("time"),
+            "id": commit_info.get("id", "unknown"),
+            "author": commit_info.get("author_name", "unknown"),
+            "message": commit_info.get("message", "unknown"),
+        },
+        "date": int(time.time() * 1000),  # Current time in ms
+        "tool": "casbin",
+        "benchmarks": [],
+    }
+
+    for b in bench_data.get("benchmarks", []):
+        name = b["name"]
+        # Name is already normalized in format_benchmark_data.py
+
+        # value is already in ns/op in format_benchmark_data.py
+        value = b["stats"]["mean"]
+
+        entry["benchmarks"].append({"name": name, "unit": "ns/op", "value": value})
+
+    # Append to entries
+    group_name = "Casbin"
+    if group_name not in js_data["entries"]:
+        js_data["entries"][group_name] = []
+
+    js_data["entries"][group_name].append(entry)
+    js_data["lastUpdate"] = int(time.time() * 1000)
+
+    save_data_js(data_js_file, js_data)
+    print(f"Updated {data_js_file} with {len(entry['benchmarks'])} benchmarks.")
+
+
+if __name__ == "__main__":
+    main()

.github/workflows/benchmark-push.yml

@@ -0,0 +1,58 @@
+name: Push Benchmark Data
+
+on:
+  push:
+    branches:
+      - master
+
+jobs:
+  benchmark:
+    name: Run Benchmark & Push Data
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout repo
+        uses: actions/checkout@v4
+
+      - name: Set up Go
+        uses: actions/setup-go@v5
+        with:
+          go-version: '1.21'
+
+      - name: Run Benchmark
+        run: |
+          go test -bench '.' -benchmem ./... > benchmark_result.txt
+
+      - name: Checkout Data Repo
+        uses: actions/checkout@v4
+        with:
+          repository: casbin/casbin-benchmark-data
+          token: ${{ secrets.CASBIN_BENCHMARK_DATA_TOKEN }}
+          path: benchmark-data
+
+      - name: Format Benchmark Data
+        env:
+          COMMIT_AUTHOR_EMAIL: ${{ github.event.head_commit.author.email }}
+          COMMIT_AUTHOR_NAME: ${{ github.event.head_commit.author.name }}
+          COMMIT_AUTHOR_USERNAME: ${{ github.event.head_commit.author.username }}
+          COMMIT_COMMITTER_EMAIL: ${{ github.event.head_commit.committer.email }}
+          COMMIT_COMMITTER_NAME: ${{ github.event.head_commit.committer.name }}
+          COMMIT_COMMITTER_USERNAME: ${{ github.event.head_commit.committer.username }}
+          COMMIT_MESSAGE: ${{ github.event.head_commit.message }}
+          COMMIT_TIMESTAMP: ${{ github.event.head_commit.timestamp }}
+          COMMIT_URL: ${{ github.event.head_commit.url }}
+          COMMIT_ID: ${{ github.event.head_commit.id }}
+        run: |
+          python .github/scripts/format_benchmark_data.py benchmark_result.txt formatted_result.json
+
+      - name: Push Benchmark Result
+        working-directory: benchmark-data
+        run: |
+          mkdir -p casbin
+          cp ../formatted_result.json casbin/benchmark-${{ github.sha }}.json
+
+          git config user.name "github-actions[bot]"
+          git config user.email "github-actions[bot]@users.noreply.github.com"
+
+          git add casbin/benchmark-${{ github.sha }}.json
+          git commit -m "Add benchmark result for casbin commit ${{ github.sha }}"
+          git push

.gitignore

@@ -28,3 +28,4 @@ _testmain.go
 
 # vendor files
 vendor
+yml-test/