Skip to content

feat: dynamic function budget for daily-function-namer (replace fixed 3-file limit) #22151

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

Merged
pelikhan merged 3 commits into from
Mar 21, 2026
Merged

feat: dynamic function budget for daily-function-namer (replace fixed 3-file limit) #22151

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
4c81159
Initial plan
Copilot Mar 21, 2026
65212cc
feat: replace fixed 3-file limit with dynamic function budget in dail…
Copilot Mar 21, 2026
d3f164e
feat: use bash script to compute file selection for dynamic function …
Copilot Mar 21, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 5 additions & 17 deletions .github/workflows/daily-function-namer.lock.yml
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

144 changes: 82 additions & 62 deletions .github/workflows/daily-function-namer.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
---
name: Daily Go Function Namer
description: Analyzes up to 3 Go files daily using Serena to extract function names and suggest renames that improve agent discoverability, using round-robin via cache-memory
description: Analyzes Go files daily using Serena to extract function names and suggest renames that improve agent discoverability, using round-robin via cache-memory with a dynamic function budget
on:
schedule: daily
workflow_dispatch:
Expand Down Expand Up @@ -30,8 +30,7 @@ tools:
cache-memory: true
github:
toolsets: [default, issues]
bash:
- "find pkg -name '*.go' ! -name '*_test.go' -type f | sort"
bash: true
Copy link

Copilot AI Mar 21, 2026

Choose a reason for hiding this comment

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

Setting tools.bash: true grants unrestricted Bash execution for the agent. If least-privilege is desired, keep an explicit allowlist of the specific commands needed by the selection script (e.g., python3, find, sort, grep, seq, etc.) instead of full Bash.

Suggested change
bash: true
bash:
allowed-commands:
- python3
- find
- sort
- grep
- seq

Copilot uses AI. Check for mistakes.

timeout-minutes: 30
strict: true
Expand All @@ -43,7 +42,7 @@ You are an AI agent that analyzes Go functions daily to improve their names for

## Mission

Each day, analyze up to **3 Go source files** using round-robin rotation across all non-test Go files in `pkg/`. For each file:
Each day, analyze up to **6 Go source files** using round-robin rotation across all non-test Go files in `pkg/`, driven by a **dynamic function budget** of 25 functions per run. For each file:

1. Extract all function and method names using Serena
2. Evaluate each name's clarity and intent
Expand All @@ -57,52 +56,71 @@ Each day, analyze up to **3 Go source files** using round-robin rotation across
- **Workspace**: ${{ github.workspace }}
- **Cache**: `/tmp/gh-aw/cache-memory/`

## Step 1: Load Round-Robin State from Cache
## Step 1: Compute File Selection with Code

Read the current rotation position from cache:
Run this script to load the round-robin state, enumerate all Go files, and compute which files to analyze this run using the dynamic function budget:

```bash
cat /tmp/gh-aw/cache-memory/function-namer-state.json
# Load last_index from cache (default 0 if cache absent/empty)
LAST_INDEX=$(python3 -c "
import sys, json, os
p = '/tmp/gh-aw/cache-memory/function-namer-state.json'
if os.path.exists(p):
try:
d = json.load(open(p))
print(d.get('last_index', 0))
except Exception:
print(0)
else:
print(0)
")

# Enumerate all non-test Go source files in sorted order
mapfile -t ALL_FILES < <(find pkg -name '*.go' ! -name '*_test.go' -type f | sort)
TOTAL=${#ALL_FILES[@]}

echo "total_files=${TOTAL}"
echo "last_index=${LAST_INDEX}"

# Greedy selection: up to 8 candidates, budget=25 functions, cap=6 files
SELECTED=()
TOTAL_FNS=0
BUDGET=25
CAP=6

for i in $(seq 0 7); do
if [ ${#SELECTED[@]} -ge $CAP ]; then
break
fi
idx=$(( (LAST_INDEX + i) % TOTAL ))
f="${ALL_FILES[$idx]}"
COUNT=$(grep -c "^func " "$f" 2>/dev/null || echo 0)
Copy link

Copilot AI Mar 21, 2026

Choose a reason for hiding this comment

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

The function-count command substitution can break for files with zero matches. grep -c prints 0 but exits with status 1 when no lines match, so grep -c ... || echo 0 produces two zeros, making COUNT non-numeric (e.g., "0 0") and causing the arithmetic update of TOTAL_FNS to fail. Adjust the pipeline so it always captures exactly one integer (e.g., ignore grep’s exit code without appending another 0, or normalize COUNT after the command).

Suggested change
COUNT=$(grep -c "^func " "$f" 2>/dev/null || echo 0)
COUNT=$(grep -c "^func " "$f" 2>/dev/null || true)
COUNT=${COUNT:-0}

Copilot uses AI. Check for mistakes.
SELECTED+=("$f")
TOTAL_FNS=$((TOTAL_FNS + COUNT))
if [ $TOTAL_FNS -ge $BUDGET ]; then
break
fi
done

SELECTED_COUNT=${#SELECTED[@]}
NEW_INDEX=$(( (LAST_INDEX + SELECTED_COUNT) % TOTAL ))

echo "selected_count=${SELECTED_COUNT}"
echo "new_last_index=${NEW_INDEX}"
echo "total_functions_approx=${TOTAL_FNS}"
echo "--- selected files ---"
printf '%s\n' "${SELECTED[@]}"
```

Expected format:
The script outputs:
- `selected_count` — number of files to analyze this run (1–6)
- `new_last_index` — value to write back to cache after the run
- `total_functions_approx` — estimated function count across selected files
- The list of selected file paths (one per line, after `--- selected files ---`)

```json
{
"last_index": 0,
"analyzed_files": [
{"file": "pkg/workflow/compiler.go", "analyzed_at": "2026-03-12"}
]
}
```

All file paths are relative to the repository root (e.g., `pkg/workflow/compiler.go`),
matching the output of the `find pkg` command in Step 3.

If the cache file does not exist or is empty, start fresh with `last_index = 0` and an
empty `analyzed_files` list.

## Step 2: Get All Go Files

Enumerate all non-test Go source files in sorted order:
Use these values directly for the rest of the workflow. Do **not** re-derive or re-compute them manually.

```bash
find pkg -name '*.go' ! -name '*_test.go' -type f | sort
```

Record the total file count for wrap-around calculations.

## Step 3: Select the Next 3 Files

Using `last_index` from the cache:

- Select files at positions `last_index`, `last_index + 1`, `last_index + 2`
- Wrap around using modulo: `index % total_files`
- Example: If there are 50 files and `last_index` is 49, select indices 49, 0, 1

The new `last_index` for the next run is `(last_index + 3) % total_files`.

## Step 4: Activate Serena
## Step 2: Activate Serena

Activate the Serena project to enable Go semantic analysis:

Expand All @@ -111,11 +129,11 @@ Tool: activate_project
Args: { "path": "${{ github.workspace }}" }
```

## Step 5: Analyze Each File with Serena
## Step 3: Analyze Each Selected File with Serena

For each of the 3 selected files, perform a full function name analysis.
For each of the selected files output by Step 1, perform a full function name analysis.

### 5.1 Get All Symbols
### 3.1 Get All Symbols

```
Tool: get_symbols_overview
Expand All @@ -124,9 +142,9 @@ Args: { "file_path": "<relative/path/to/file.go>" }

This returns all functions, methods, and types defined in the file.

### 5.2 Read Function Implementations
### 3.2 Read Function Implementations

For each function identified in 6.1, read enough of the implementation to understand its behavior:
For each function identified in 3.1, read enough of the implementation to understand its behavior:

```
Tool: read_file
Expand All @@ -139,7 +157,7 @@ For small files you may read the entire file:
cat <path/to/file.go>
```

### 5.3 Evaluate Function Names
### 3.3 Evaluate Function Names

For each function, assess its name against these criteria:

Expand All @@ -156,7 +174,7 @@ For each function, assess its name against these criteria:
- Constructors following Go convention: `NewCompiler()`, `NewMCPConfig()`
- Short unexported names used as closures or immediately-invoked helpers

### 5.4 Propose Renames
### 3.4 Propose Renames

For each function that would benefit from a clearer name:

Expand All @@ -180,9 +198,9 @@ Args: { "symbol_name": "<currentName>", "file_path": "pkg/..." }

**Only suggest renames where the improvement is clear and meaningful.** Quality over quantity — two well-justified suggestions are better than ten marginal ones.

## Step 6: Update Cache State
## Step 4: Update Cache State

After completing the analysis, save the updated round-robin position. Use a filesystem-safe timestamp format (`YYYY-MM-DD` is fine for daily granularity):
After completing the analysis, save the updated round-robin position. Use the `new_last_index` value from Step 1 and a filesystem-safe timestamp (`YYYY-MM-DD`):

```bash
cat > /tmp/gh-aw/cache-memory/function-namer-state.json << 'CACHE_EOF'
Expand All @@ -191,41 +209,43 @@ cat > /tmp/gh-aw/cache-memory/function-namer-state.json << 'CACHE_EOF'
"analyzed_files": [
<previous entries, pruned to last 90>,
{"file": "pkg/workflow/compiler.go", "analyzed_at": "2026-03-13"},
{"file": "pkg/workflow/cache.go", "analyzed_at": "2026-03-13"},
{"file": "pkg/workflow/mcp_renderer.go", "analyzed_at": "2026-03-13"}
{"file": "pkg/workflow/cache.go", "analyzed_at": "2026-03-13"}
]
}
CACHE_EOF
```

Where `<new_index>` is the `new_last_index` value output by Step 1, and the `analyzed_files` list contains one entry per file actually analyzed.

Use relative paths (e.g., `pkg/workflow/compiler.go`) matching the output of the `find pkg` command.

Prune `analyzed_files` to the most recent 90 entries to prevent unbounded growth.

## Step 7: Create Issue with Agentic Plan
## Step 5: Create Issue with Agentic Plan

If any rename suggestions were found across the 3 files, create a GitHub issue.
If any rename suggestions were found across the analyzed files, create a GitHub issue.

If **no improvements were found**, emit `noop` and exit:

```json
{"noop": {"message": "No rename suggestions found for <file1>, <file2>, <file3>. All analyzed functions have clear, descriptive names."}}
{"noop": {"message": "No rename suggestions found for <file1>, …, <fileN>. All analyzed functions have clear, descriptive names."}}
```

Otherwise, create an issue with this structure:

---

**Title**: `Go function rename plan: <basename1>, <basename2>, <basename3>` (e.g., `Go function rename plan: compiler.go, cache.go, mcp_renderer.go`)
**Title**: `Go function rename plan: <basename1>, <basename2>, …` (e.g., `Go function rename plan: compiler.go, cache.go, mcp_renderer.go`)

**Body**:

```markdown
# 🏷️ Go Function Rename Plan

**Files Analyzed**: `<file1>`, `<file2>`, `<file3>`
**Files Analyzed**: `<file1>`, `<file2>`, …
**Analysis Date**: <YYYY-MM-DD>
**Round-Robin Position**: files <start_index>–<end_index> of <total> total
**Function Budget**: <total_functions_analyzed> functions across <selected_count> files
Copy link

Copilot AI Mar 21, 2026

Choose a reason for hiding this comment

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

The issue template placeholder Function Budget: <total_functions_analyzed> ... doesn’t correspond to any named output from Step 1 (which emits total_functions_approx). Consider aligning the placeholder/name with the script output (or clarify that it should be the approximate count) to avoid confusion when the agent fills in the issue body.

Suggested change
**Function Budget**: <total_functions_analyzed> functions across <selected_count> files
**Function Budget**: ~<total_functions_approx> functions across <selected_count> files

Copilot uses AI. Check for mistakes.

### Why This Matters

Expand All @@ -249,9 +269,9 @@ that an agent will find the right function instead of reimplementing existing lo

<!-- Same structure, or: "No renames needed for this file." -->

#### `<file3>`
#### `<fileN>`

<!-- Same structure, or: "No renames needed for this file." -->
<!-- Repeat for each analyzed file. -->

---

Expand Down Expand Up @@ -373,5 +393,5 @@ Only include a rename suggestion if you are confident it would measurably improv
**Important**: If no action is needed after completing your analysis, you **MUST** call the `noop` safe-output tool. Failing to call any safe-output tool is the most common cause of workflow failures.

```json
{"noop": {"message": "No rename suggestions found for <file1>, <file2>, <file3>. All analyzed functions already have clear, descriptive names that support agent discoverability."}}
{"noop": {"message": "No rename suggestions found for <file1>, …, <fileN>. All analyzed functions already have clear, descriptive names that support agent discoverability."}}
```
Loading