run_protocol_tests.sh

#!/bin/bash

# A simple test runner for the Agent Feedback Protocol v1.2

# --- Colors for output ---
RED='\033[0;31m'
GREEN='\033[0;32m'
YELLOW='\033[1;33m'
NC='\033[0m' # No Color

# --- Test Fixture Paths ---
# Assuming the script is run from the project root
FIXTURE_PATH="tests/fixtures"
REQUEST_T1="$FIXTURE_PATH/request_t1.json"
REQUEST_T2="$FIXTURE_PATH/request_t2.json"
REQUEST_MALFORMED="$FIXTURE_PATH/request_malformed.txt"
REQUEST_INCOMPLETE="$FIXTURE_PATH/request_incomplete.json"

# --- Temp file for output ---
OUTPUT_FILE=$(mktemp)

# --- Cleanup function ---
cleanup() {
  rm -f "$OUTPUT_FILE"
}
trap cleanup EXIT

# --- Helper for running a test ---
# Usage: run_test "Test Name" "command_to_run" "jq_filter" "expected_jq_result"
run_test() {
  local test_name="$1"
  local command="$2"
  local jq_filter="$3"
  local expected_result="$4"

  echo -e "\n${YELLOW}Running Test: $test_name${NC}"

  # Execute the command and capture stdout to a file, stderr is ignored for now
  eval "$command" > "$OUTPUT_FILE" 2>/dev/null
  local exit_code=$?

  # Check if jq is installed
  if ! command -v jq &> /dev/null;
  then
      echo -e "${RED}FAIL: 'jq' is not installed. Cannot verify JSON output.${NC}"
      return 1
  fi

  # For tests that are expected to succeed on the command level
  if [[ "$test_name" != *"Invalid Request"* ]];
  then
    if [ $exit_code -ne 0 ];
    then
        echo -e "${RED}FAIL: Command failed unexpectedly with exit code $exit_code${NC}"
        cat "$OUTPUT_FILE"
        return 1
    fi
  fi

  # Verify the output with jq
  local actual_result=$(jq -r "$jq_filter" "$OUTPUT_FILE")

  if [ "$actual_result" == "$expected_result" ];
  then
    echo -e "${GREEN}PASS${NC}"
    return 0
  else
    echo -e "${RED}FAIL: Assertion failed.${NC}"
    echo "  Expected: $expected_result"
    echo "  Actual:   $actual_result"
    echo "  Full output:"
    cat "$OUTPUT_FILE"
    return 1
  fi
}

# --- Test Cases ---

test_case_1() {
  run_test \
    "Test Case 1: New Session Initiation" \
    "cat $REQUEST_T1 | opencode run --agent qualitative-reviewer" \
    ".protocol_version" \
    "1.2"
}

test_case_2() {
  # This test uses a hardcoded session_id for simplicity.
  # A more advanced script would capture the session_id from test_case_1.
  run_test \
    "Test Case 2: Follow-up Session with Feedback Selection" \
    "cat $REQUEST_T2 | opencode run --agent qualitative-reviewer --session session-123" \
    ".applied_feedback_ack.items[0].processing_status" \
    "acknowledged"
}

test_case_3() {
    run_test \
    "Test Case 3: Invalid Request (Malformed JSON)" \
    "cat $REQUEST_MALFORMED | opencode run --agent qualitative-reviewer" \
    ".error.code" \
    "INVALID_REQUEST"
}

test_case_4() {
    run_test \
    "Test Case 4: Invalid Request (Missing Required Field)" \
    "cat $REQUEST_INCOMPLETE | opencode run --agent qualitative-reviewer" \
    ".error.code" \
    "INVALID_REQUEST"
}


# --- Main execution ---
main() {
  echo "Starting Agent Feedback Protocol v1.2 Test Suite..."
  local overall_status=0

  test_case_1 || overall_status=1
  # SKIPPING: Test Case 2 is disabled due to a known bug in the opencode CLI's --session flag handling.
  # test_case_2 || overall_status=1
  test_case_3 || overall_status=1
  test_case_4 || overall_status=1

  echo ""
  if [ $overall_status -eq 0 ];
  then
    echo -e "${GREEN}All tests seem to have passed!${NC}"
  else
    echo -e "${RED}Some tests failed.${NC}"
  fi

  exit $overall_status
}

main