ros/claude-task-master
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Refactor: Improve MCP logging, update E2E & tests

Browse Source 
Refactors MCP server logging and updates testing infrastructure.

- MCP Server:

  - Replaced manual logger wrappers with centralized `createLogWrapper` utility.

  - Updated direct function calls to use `{ session, mcpLog }` context.

  - Removed deprecated `model` parameter from analyze, expand-all, expand-task tools.

  - Adjusted MCP tool import paths and parameter descriptions.

- Documentation:

  - Modified `docs/configuration.md`.

  - Modified `docs/tutorial.md`.

- Testing:

  - E2E Script (`run_e2e.sh`):

    - Removed `set -e`.

    - Added LLM analysis function (`analyze_log_with_llm`) & integration.

    - Adjusted test run directory creation timing.

    - Added debug echo statements.

  - Deleted Unit Tests: Removed `ai-client-factory.test.js`, `ai-client-utils.test.js`, `ai-services.test.js`.

  - Modified Fixtures: Updated `scripts/task-complexity-report.json`.

- Dev Scripts:

  - Modified `scripts/dev.js`.
This commit is contained in:
Eyal Toledano
2025-04-28 14:38:01 -04:00
parent 6d4471fcb5
commit 4ac01f33c4
37 changed files with 687 additions and 1736 deletions
Download Patch File Download Diff File Expand all files Collapse all files

248
tests/e2e/run_e2e.sh
View File

@@ -1,7 +1,5 @@
#!/bin/bash
# Exit immediately if a command exits with a non-zero status.
set -e
# Treat unset variables as an error when substituting.
set -u
# Prevent errors in pipelines from being masked.
@@ -33,6 +31,11 @@ mkdir -p "$LOG_DIR"
TIMESTAMP=$(date +"%Y%m%d_%H%M%S")
LOG_FILE="$LOG_DIR/e2e_run_$TIMESTAMP.log"
# Define and create the test run directory *before* the main pipe
mkdir -p "$BASE_TEST_DIR" # Ensure base exists first
TEST_RUN_DIR="$BASE_TEST_DIR/run_$TIMESTAMP"
mkdir -p "$TEST_RUN_DIR"
# Echo starting message to the original terminal BEFORE the main piped block
echo "Starting E2E test. Output will be shown here and saved to: $LOG_FILE"
echo "Running from directory: $(pwd)"
@@ -82,6 +85,125 @@ overall_start_time=$(date +%s)
echo " STEP ${test_step_count}: [$(_get_elapsed_time_for_log)] $(date +"%Y-%m-%d %H:%M:%S") $1"
echo "============================================="
}
analyze_log_with_llm() {
local log_file="$1"
local provider_summary_log="provider_add_task_summary.log" # File summarizing provider test outcomes
local api_key=""
local api_endpoint="https://api.anthropic.com/v1/messages"
local api_key_name="CLAUDE_API_KEY"
echo "" # Add a newline before analysis starts
log_info "Attempting LLM analysis of log: $log_file"
# Check for jq and curl
if ! command -v jq &> /dev/null; then
log_error "LLM Analysis requires 'jq'. Skipping analysis."
return 1
fi
if ! command -v curl &> /dev/null; then
log_error "LLM Analysis requires 'curl'. Skipping analysis."
return 1
fi
# Check for API Key in the TEST_RUN_DIR/.env (copied earlier)
if [ -f ".env" ]; then
# Using grep and sed for better handling of potential quotes/spaces
api_key=$(grep "^${api_key_name}=" .env | sed -e "s/^${api_key_name}=//" -e 's/^[[:space:]"]*//' -e 's/[[:space:]"]*$//')
fi
if [ -z "$api_key" ]; then
log_error "${api_key_name} not found or empty in .env file in the test run directory ($(pwd)/.env). Skipping LLM analysis."
return 1
fi
if [ ! -f "$log_file" ]; then
log_error "Log file not found: $log_file. Skipping LLM analysis."
return 1
fi
log_info "Reading log file content..."
local log_content
# Read entire file, handle potential errors
log_content=$(cat "$log_file") || {
log_error "Failed to read log file: $log_file. Skipping LLM analysis."
return 1
}
# Prepare the prompt
# Using printf with %s for the log content is generally safer than direct variable expansion
local prompt_template='Analyze the following E2E test log for the task-master tool. The log contains output from various '\''task-master'\'' commands executed sequentially.\n\nYour goal is to:\n1. Verify if the key E2E steps completed successfully based on the log messages (e.g., init, parse PRD, list tasks, analyze complexity, expand task, set status, manage models, add/remove dependencies, add/update/remove tasks/subtasks, generate files).\n2. **Specifically analyze the Multi-Provider Add-Task Test Sequence:**\n a. Identify which providers were tested for `add-task`. Look for log steps like "Testing Add-Task with Provider: ..." and the summary log `'"$provider_summary_log"'`.\n b. For each tested provider, determine if `add-task` succeeded or failed. Note the created task ID if successful.\n c. Review the corresponding `add_task_show_output_<provider>_id_<id>.log` file (if created) for each successful `add-task` execution.\n d. **Compare the quality and completeness** of the task generated by each successful provider based on their `show` output. Assign a score (e.g., 1-10, 10 being best) based on relevance to the prompt, detail level, and correctness.\n e. Note any providers where `add-task` failed or where the task ID could not be extracted.\n3. Identify any general explicit "[ERROR]" messages or stack traces throughout the *entire* log.\n4. Identify any potential warnings or unusual output that might indicate a problem even if not marked as an explicit error.\n5. Provide an overall assessment of the test run'\''s health based *only* on the log content.\n\nReturn your analysis **strictly** in the following JSON format. Do not include any text outside of the JSON structure:\n\n{\n "overall_status": "Success|Failure|Warning",\n "verified_steps": [ "Initialization", "PRD Parsing", /* ...other general steps observed... */ ],\n "provider_add_task_comparison": {\n "prompt_used": "... (extract from log if possible or state 'standard auth prompt') ...",\n "provider_results": {\n "anthropic": { "status": "Success|Failure|ID_Extraction_Failed|Set_Model_Failed", "task_id": "...", "score": "X/10 | N/A", "notes": "..." },\n "openai": { "status": "Success|Failure|...", "task_id": "...", "score": "X/10 | N/A", "notes": "..." },\n /* ... include all tested providers ... */\n },\n "comparison_summary": "Brief overall comparison of generated tasks..."\n },\n "detected_issues": [ { "severity": "Error|Warning|Anomaly", "description": "...", "log_context": "[Optional, short snippet from log near the issue]" } ],\n "llm_summary_points": [ "Overall summary point 1", "Provider comparison highlight", "Any major issues noted" ]\n}\n\nHere is the main log content:\n\n%s'
local full_prompt
printf -v full_prompt "$prompt_template" "$log_content"
# Construct the JSON payload for Claude Messages API
# Using jq for robust JSON construction
local payload
payload=$(jq -n --arg prompt "$full_prompt" '{
"model": "claude-3-7-sonnet-20250219",
"max_tokens": 10000,
"messages": [
{"role": "user", "content": $prompt}
],
"temperature": 0.0
}') || {
log_error "Failed to create JSON payload using jq."
return 1
}
log_info "Sending request to LLM API endpoint: $api_endpoint ..."
local response_raw response_http_code response_body
# Capture body and HTTP status code separately
response_raw=$(curl -s -w "\nHTTP_STATUS_CODE:%{http_code}" -X POST "$api_endpoint" \
-H "Content-Type: application/json" \
-H "x-api-key: $api_key" \
-H "anthropic-version: 2023-06-01" \
--data "$payload")
# Extract status code and body
response_http_code=$(echo "$response_raw" | grep '^HTTP_STATUS_CODE:' | sed 's/HTTP_STATUS_CODE://')
response_body=$(echo "$response_raw" | sed '$d') # Remove last line (status code)
if [ "$response_http_code" != "200" ]; then
log_error "LLM API call failed with HTTP status $response_http_code."
log_error "Response Body: $response_body"
return 1
fi
if [ -z "$response_body" ]; then
log_error "LLM API call returned empty response body."
return 1
fi
log_info "Received LLM response (HTTP 200). Parsing analysis JSON..."
# Extract the analysis JSON string from the API response (adjust jq path if needed)
local analysis_json_string
analysis_json_string=$(echo "$response_body" | jq -r '.content[0].text' 2>/dev/null) # Assumes Messages API structure
if [ -z "$analysis_json_string" ]; then
log_error "Failed to extract 'content[0].text' from LLM response JSON."
log_error "Full API response body: $response_body"
return 1
fi
# Validate and pretty-print the extracted JSON
if ! echo "$analysis_json_string" | jq -e . > /dev/null 2>&1; then
log_error "Extracted content from LLM is not valid JSON."
log_error "Raw extracted content: $analysis_json_string"
return 1
fi
log_success "LLM analysis completed successfully."
echo ""
echo "--- LLM Analysis ---"
# Pretty print the JSON analysis
echo "$analysis_json_string" | jq '.'
echo "--------------------"
return 0
}
# ---
# --- Test Setup (Output to tee) ---
@@ -95,12 +217,9 @@ overall_start_time=$(date +%s)
exit 1
fi
mkdir -p "$BASE_TEST_DIR"
log_info "Ensured base test directory exists: $BASE_TEST_DIR"
TEST_RUN_DIR="$BASE_TEST_DIR/run_$TIMESTAMP"
mkdir -p "$TEST_RUN_DIR"
log_info "Created test run directory: $TEST_RUN_DIR"
log_info "Using test run directory (created earlier): $TEST_RUN_DIR"
# Check if source .env file exists
if [ ! -f "$MAIN_ENV_FILE" ]; then
@@ -209,8 +328,103 @@ overall_start_time=$(date +%s)
log_step "Checking final model configuration"
task-master models > models_final_config.log
log_success "Final model config saved to models_final_config.log"
log_step "Resetting main model to default (Claude Sonnet) before provider tests"
task-master models --set-main claude-3-7-sonnet-20250219
log_success "Main model reset to claude-3-7-sonnet-20250219."
# === End Model Commands Test ===
# === Multi-Provider Add-Task Test ===
log_step "Starting Multi-Provider Add-Task Test Sequence"
# Define providers, models, and flags
# Array order matters: providers[i] corresponds to models[i] and flags[i]
declare -a providers=("anthropic" "openai" "google" "perplexity" "xai" "openrouter")
declare -a models=(
"claude-3-7-sonnet-20250219"
"gpt-4o"
"gemini-2.5-pro-exp-03-25"
"sonar-pro"
"grok-3"
"anthropic/claude-3.7-sonnet" # OpenRouter uses Claude 3.7
)
# Flags: Add provider-specific flags here, e.g., --openrouter. Use empty string if none.
declare -a flags=("" "" "" "" "" "--openrouter")
# Consistent prompt for all providers
add_task_prompt="Create a task to implement user authentication using OAuth 2.0 with Google as the provider. Include steps for registering the app, handling the callback, and storing user sessions."
log_info "Using consistent prompt for add-task tests: \"$add_task_prompt\""
for i in "${!providers[@]}"; do
provider="${providers[$i]}"
model="${models[$i]}"
flag="${flags[$i]}"
log_step "Testing Add-Task with Provider: $provider (Model: $model)"
# 1. Set the main model for this provider
log_info "Setting main model to $model for $provider ${flag:+using flag $flag}..."
set_model_cmd="task-master models --set-main \"$model\" $flag"
echo "Executing: $set_model_cmd"
if eval $set_model_cmd; then
log_success "Successfully set main model for $provider."
else
log_error "Failed to set main model for $provider. Skipping add-task for this provider."
# Optionally save failure info here if needed for LLM analysis
echo "Provider $provider set-main FAILED" >> provider_add_task_summary.log
continue # Skip to the next provider
fi
# 2. Run add-task
log_info "Running add-task with prompt..."
add_task_output_file="add_task_raw_output_${provider}.log"
# Run add-task and capture ALL output (stdout & stderr) to a file AND a variable
add_task_cmd_output=$(task-master add-task --prompt "$add_task_prompt" 2>&1 | tee "$add_task_output_file")
add_task_exit_code=${PIPESTATUS[0]}
# 3. Check for success and extract task ID
new_task_id=""
if [ $add_task_exit_code -eq 0 ] && echo "$add_task_cmd_output" | grep -q "Successfully added task with ID:"; then
# Attempt to extract the ID (adjust grep/sed/awk as needed based on actual output format)
new_task_id=$(echo "$add_task_cmd_output" | grep "Successfully added task with ID:" | sed 's/.*Successfully added task with ID: \([0-9.]\+\).*/\1/')
if [ -n "$new_task_id" ]; then
log_success "Add-task succeeded for $provider. New task ID: $new_task_id"
echo "Provider $provider add-task SUCCESS (ID: $new_task_id)" >> provider_add_task_summary.log
else
# Succeeded but couldn't parse ID - treat as warning/anomaly
log_error "Add-task command succeeded for $provider, but failed to extract task ID from output."
echo "Provider $provider add-task SUCCESS (ID extraction FAILED)" >> provider_add_task_summary.log
new_task_id="UNKNOWN_ID_EXTRACTION_FAILED"
fi
else
log_error "Add-task command failed for $provider (Exit Code: $add_task_exit_code). See $add_task_output_file for details."
echo "Provider $provider add-task FAILED (Exit Code: $add_task_exit_code)" >> provider_add_task_summary.log
new_task_id="FAILED"
fi
# 4. Run task show if ID was obtained (even if extraction failed, use placeholder)
if [ "$new_task_id" != "FAILED" ] && [ "$new_task_id" != "UNKNOWN_ID_EXTRACTION_FAILED" ]; then
log_info "Running task show for new task ID: $new_task_id"
show_output_file="add_task_show_output_${provider}_id_${new_task_id}.log"
if task-master show "$new_task_id" > "$show_output_file"; then
log_success "Task show output saved to $show_output_file"
else
log_error "task show command failed for ID $new_task_id. Check log."
# Still keep the file, it might contain error output
fi
elif [ "$new_task_id" == "UNKNOWN_ID_EXTRACTION_FAILED" ]; then
log_info "Skipping task show for $provider due to ID extraction failure."
else
log_info "Skipping task show for $provider due to add-task failure."
fi
done # End of provider loop
log_step "Finished Multi-Provider Add-Task Test Sequence"
echo "Provider add-task summary log available at: provider_add_task_summary.log"
# === End Multi-Provider Add-Task Test ===
log_step "Listing tasks again (final)"
task-master list --with-subtasks > task_list_final.log
log_success "Final task list saved to task_list_final.log"
@@ -386,4 +600,26 @@ else
fi
echo "-------------------------"
# --- Attempt LLM Analysis ---
echo "DEBUG: Entering LLM Analysis section..."
# Run this *after* the main execution block and tee pipe finish writing the log file
# It will read the completed log file and append its output to the terminal (and the log via subsequent writes if tee is still active, though it shouldn't be)
# Change directory back into the test run dir where .env is located
if [ -d "$TEST_RUN_DIR" ]; then
echo "DEBUG: Found TEST_RUN_DIR: $TEST_RUN_DIR. Attempting cd..."
cd "$TEST_RUN_DIR"
echo "DEBUG: Changed directory to $(pwd). Calling analyze_log_with_llm..."
analyze_log_with_llm "$LOG_FILE"
echo "DEBUG: analyze_log_with_llm function call finished."
# Optional: cd back again if needed, though script is ending
# cd "$ORIGINAL_DIR"
else
# Use log_error format even outside the pipe for consistency
current_time_for_error=$(date +%s)
elapsed_seconds_for_error=$((current_time_for_error - overall_start_time)) # Use overall start time
formatted_duration_for_error=$(_format_duration "$elapsed_seconds_for_error")
echo "[ERROR] [$formatted_duration_for_error] $(date +"%Y-%m-%d %H:%M:%S") Test run directory $TEST_RUN_DIR not found. Cannot perform LLM analysis." >&2
fi
echo "DEBUG: Reached end of script before final exit."
exit $EXIT_CODE # Exit with the status of the main script block