looper-dashboard 0.1.0__tar.gz

looper_dashboard-0.1.0/PKG-INFO

@@ -0,0 +1,198 @@
+Metadata-Version: 2.4
+Name: looper-dashboard
+Version: 0.1.0
+Summary: Generate daily CI dashboards for any QE repo on LooperPro (Karate, Playwright, Monocart)
+License: MIT
+Keywords: karate,playwright,monocart,looperpro,ci,dashboard
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+Requires-Dist: pyyaml>=6.0
+Requires-Dist: requests>=2.31
+
+# looper-dashboard
+
+Generate daily CI dashboards for any QE repository on the [LooperPro](https://looperpro.prod.walmart.com) CI platform.
+
+Supports three test suite types:
+
+| Suite | Report format | Default history |
+|---|---|---|
+| `karate` | `karate-summary-json.txt` + per-feature detail | 14 days |
+| `playwright` | `pw-summary-report.json` | 14 days |
+| `playwright-monocart` | `monocart-report/index.html` (zlib-compressed) | 30 days |
+
+---
+
+## Requirements
+
+| Requirement | Why |
+|---|---|
+| Python 3.11+ | Runtime |
+| `mcp-cli` (optional) | Auto-refreshes LooperPro auth tokens |
+| `code-puppy` (optional) | Powers the AI failure analysis step |
+
+Authentication is read from `~/.mcp-cli/tokens.json` or `~/.code_puppy/puppy.cfg` automatically — no extra config needed.
+
+---
+
+## Installation
+
+```bash
+pip install looper-dashboard
+```
+
+Or from the repo:
+
+```bash
+pip install ./looper-dashboard
+```
+
+---
+
+## Quick Start
+
+### 1. Scaffold a config
+
+```bash
+# Creates daily_flows.yml in the current directory
+looper-dashboard --init karate
+looper-dashboard --init playwright
+looper-dashboard --init playwright-monocart
+```
+
+Edit the generated file to add your LooperPro job names and flow names.
+
+### 2. Generate the dashboard
+
+```bash
+looper-dashboard --suite karate --config daily_flows.yml --open
+looper-dashboard --suite playwright --config daily_flows.yml --open
+looper-dashboard --suite playwright-monocart --config daily_flows.yml --open
+```
+
+### 3. Override the repo SCM URL
+
+```bash
+looper-dashboard --suite karate \
+  --config daily_flows.yml \
+  --scm-url https://gecgithub01.walmart.com/my-org/my-repo.git \
+  --open
+```
+
+---
+
+## `daily_flows.yml` format
+
+### Karate / Playwright
+
+```yaml
+# SCM URL of the repository being monitored
+scm_url: https://gecgithub01.walmart.com/my-org/my-repo.git
+
+jobs:
+
+  # LooperPro job name → list of flow names to monitor
+  my-org/my-repo-tests:
+    - my-flow-dev
+    - my-flow-stage
+    - my-flow-prod
+
+  my-org/my-repo-health-checks:
+    - health-dev
+    - health-stage
+```
+
+### Playwright-Monocart (with explicit job_ids)
+
+Some repos' jobs cannot be discovered by SCM URL in LooperPro. Add a `job_ids` section with the UUIDs from the LooperPro UI:
+
+```yaml
+scm_url: https://gecgithub01.walmart.com/my-org/my-repo.git
+
+# Explicit job IDs (look these up in the LooperPro UI)
+job_ids:
+  my-monocart-job: "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
+
+jobs:
+
+  my-monocart-job:
+    - e2e-flow-dev
+    - e2e-flow-stage
+```
+
+---
+
+## CLI Reference
+
+```
+looper-dashboard [OPTIONS]
+
+Options:
+  --suite SUITE           Test suite type: karate, playwright, playwright-monocart  [required]
+  --config PATH           Path to daily_flows.yml  [required]
+  --scm-url URL           Override scm_url from daily_flows.yml
+  --output PATH           Output HTML file path (default varies by suite)
+  --days N                Days of build history to fetch (default: 14 or 30)
+  --open                  Auto-open dashboard in browser when done
+  --no-ai                 Skip the AI failure analysis step
+  --no-details            Skip fetching test summary details (faster, status only)
+  --save-failures PATH    Path to save failures JSON
+  --analysis-json PATH    Use a pre-computed AI analysis JSON instead of running live AI
+  --init SUITE            Scaffold a starter daily_flows.yml and exit
+
+  -h, --help              Show help and exit
+```
+
+---
+
+## AI Analysis
+
+When `--no-ai` is not set, the orchestrator:
+
+1. Saves a `*-failures.json` summary of all failing flows
+2. Invokes `code-puppy` to classify each failure as one of:
+   - `service_regression` — likely a code regression in the app
+   - `environment_issue` — infra/network/auth failure
+   - `test_issue` — stale selector or bad test data
+   - `flaky` — intermittent, no clear pattern
+3. Regenerates the dashboard HTML with an AI analysis panel embedded
+
+If `code-puppy` is not installed, the AI step is skipped gracefully.
+
+---
+
+## Publishing
+
+### To public PyPI
+
+```bash
+cd looper-dashboard
+pip install hatch
+hatch build          # creates dist/looper_dashboard-0.1.0.tar.gz and .whl
+hatch publish        # prompts for PyPI API token
+```
+
+### To Walmart internal PyPI registry
+
+Add the registry to `~/.config/hatch/config.toml`:
+
+```toml
+[publish.index.repos.internal]
+url = "https://your-internal-pypi-registry/simple/"
+```
+
+Then:
+
+```bash
+hatch publish --repo internal
+```
+
+---
+
+## Onboarding a new repo
+
+1. Create a `daily_flows.yml` for the target repo (use `--init` to scaffold)
+2. Add job and flow names (find them in LooperPro or via `mcp-cli`)
+3. Run `looper-dashboard --suite <type> --config daily_flows.yml --open`
+
+No code changes required — the package is entirely configuration-driven.

looper_dashboard-0.1.0/README.md

@@ -0,0 +1,187 @@
+# looper-dashboard
+
+Generate daily CI dashboards for any QE repository on the [LooperPro](https://looperpro.prod.walmart.com) CI platform.
+
+Supports three test suite types:
+
+| Suite | Report format | Default history |
+|---|---|---|
+| `karate` | `karate-summary-json.txt` + per-feature detail | 14 days |
+| `playwright` | `pw-summary-report.json` | 14 days |
+| `playwright-monocart` | `monocart-report/index.html` (zlib-compressed) | 30 days |
+
+---
+
+## Requirements
+
+| Requirement | Why |
+|---|---|
+| Python 3.11+ | Runtime |
+| `mcp-cli` (optional) | Auto-refreshes LooperPro auth tokens |
+| `code-puppy` (optional) | Powers the AI failure analysis step |
+
+Authentication is read from `~/.mcp-cli/tokens.json` or `~/.code_puppy/puppy.cfg` automatically — no extra config needed.
+
+---
+
+## Installation
+
+```bash
+pip install looper-dashboard
+```
+
+Or from the repo:
+
+```bash
+pip install ./looper-dashboard
+```
+
+---
+
+## Quick Start
+
+### 1. Scaffold a config
+
+```bash
+# Creates daily_flows.yml in the current directory
+looper-dashboard --init karate
+looper-dashboard --init playwright
+looper-dashboard --init playwright-monocart
+```
+
+Edit the generated file to add your LooperPro job names and flow names.
+
+### 2. Generate the dashboard
+
+```bash
+looper-dashboard --suite karate --config daily_flows.yml --open
+looper-dashboard --suite playwright --config daily_flows.yml --open
+looper-dashboard --suite playwright-monocart --config daily_flows.yml --open
+```
+
+### 3. Override the repo SCM URL
+
+```bash
+looper-dashboard --suite karate \
+  --config daily_flows.yml \
+  --scm-url https://gecgithub01.walmart.com/my-org/my-repo.git \
+  --open
+```
+
+---
+
+## `daily_flows.yml` format
+
+### Karate / Playwright
+
+```yaml
+# SCM URL of the repository being monitored
+scm_url: https://gecgithub01.walmart.com/my-org/my-repo.git
+
+jobs:
+
+  # LooperPro job name → list of flow names to monitor
+  my-org/my-repo-tests:
+    - my-flow-dev
+    - my-flow-stage
+    - my-flow-prod
+
+  my-org/my-repo-health-checks:
+    - health-dev
+    - health-stage
+```
+
+### Playwright-Monocart (with explicit job_ids)
+
+Some repos' jobs cannot be discovered by SCM URL in LooperPro. Add a `job_ids` section with the UUIDs from the LooperPro UI:
+
+```yaml
+scm_url: https://gecgithub01.walmart.com/my-org/my-repo.git
+
+# Explicit job IDs (look these up in the LooperPro UI)
+job_ids:
+  my-monocart-job: "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
+
+jobs:
+
+  my-monocart-job:
+    - e2e-flow-dev
+    - e2e-flow-stage
+```
+
+---
+
+## CLI Reference
+
+```
+looper-dashboard [OPTIONS]
+
+Options:
+  --suite SUITE           Test suite type: karate, playwright, playwright-monocart  [required]
+  --config PATH           Path to daily_flows.yml  [required]
+  --scm-url URL           Override scm_url from daily_flows.yml
+  --output PATH           Output HTML file path (default varies by suite)
+  --days N                Days of build history to fetch (default: 14 or 30)
+  --open                  Auto-open dashboard in browser when done
+  --no-ai                 Skip the AI failure analysis step
+  --no-details            Skip fetching test summary details (faster, status only)
+  --save-failures PATH    Path to save failures JSON
+  --analysis-json PATH    Use a pre-computed AI analysis JSON instead of running live AI
+  --init SUITE            Scaffold a starter daily_flows.yml and exit
+
+  -h, --help              Show help and exit
+```
+
+---
+
+## AI Analysis
+
+When `--no-ai` is not set, the orchestrator:
+
+1. Saves a `*-failures.json` summary of all failing flows
+2. Invokes `code-puppy` to classify each failure as one of:
+   - `service_regression` — likely a code regression in the app
+   - `environment_issue` — infra/network/auth failure
+   - `test_issue` — stale selector or bad test data
+   - `flaky` — intermittent, no clear pattern
+3. Regenerates the dashboard HTML with an AI analysis panel embedded
+
+If `code-puppy` is not installed, the AI step is skipped gracefully.
+
+---
+
+## Publishing
+
+### To public PyPI
+
+```bash
+cd looper-dashboard
+pip install hatch
+hatch build          # creates dist/looper_dashboard-0.1.0.tar.gz and .whl
+hatch publish        # prompts for PyPI API token
+```
+
+### To Walmart internal PyPI registry
+
+Add the registry to `~/.config/hatch/config.toml`:
+
+```toml
+[publish.index.repos.internal]
+url = "https://your-internal-pypi-registry/simple/"
+```
+
+Then:
+
+```bash
+hatch publish --repo internal
+```
+
+---
+
+## Onboarding a new repo
+
+1. Create a `daily_flows.yml` for the target repo (use `--init` to scaffold)
+2. Add job and flow names (find them in LooperPro or via `mcp-cli`)
+3. Run `looper-dashboard --suite <type> --config daily_flows.yml --open`
+
+No code changes required — the package is entirely configuration-driven.

looper_dashboard-0.1.0/looper_dashboard/__init__.py

@@ -0,0 +1,3 @@
+"""looper-dashboard — daily CI dashboards for any QE repo on LooperPro."""
+
+__version__ = "0.1.0"

looper_dashboard-0.1.0/looper_dashboard/ai_analyzer.py

@@ -0,0 +1,221 @@
+"""AI failure analyzer for Playwright and Monocart dashboards.
+
+Invokes playwright-monocart-helper via the code-puppy CLI to batch-analyze
+all failing/flaky flows and return structured classifications.
+
+Usage (standalone):
+    python3 -m looper_dashboard.ai_analyzer
+    python3 -m looper_dashboard.ai_analyzer --failures failures.json --output analysis.json
+"""
+import base64
+import json
+import re
+import subprocess
+import sys
+from pathlib import Path
+
+# ---------------------------------------------------------------------------
+# Constants
+# ---------------------------------------------------------------------------
+
+_VENV_BIN = Path.home() / ".code-puppy-venv" / "bin" / "code-puppy"
+AGENT = "playwright-monocart-helper"
+TIMEOUT = 240  # seconds
+MAX_TITLES = 15  # max test titles per flow to keep token usage reasonable
+
+
+# ---------------------------------------------------------------------------
+# ANSI / OSC stripping
+# ---------------------------------------------------------------------------
+
+def _strip_ansi(text: str) -> str:
+    """Remove ANSI/VT escape sequences (including OSC) from text."""
+    text = re.sub(r'\x1b\[[0-9;]*[A-Za-z]', '', text)
+    text = re.sub(r'\x1b\].*?(?:\x07|\x1b\\)', '', text, flags=re.DOTALL)
+    text = re.sub(r'\x1b.', '', text)
+    return text
+
+
+def _extract_json_from_osc52(text: str) -> list | None:
+    """Extract and decode a base64 payload from an OSC 52 clipboard sequence."""
+    m = re.search(r']\s*52\s*;\s*c\s*;\s*([A-Za-z0-9+/=]+)', text)
+    if not m:
+        return None
+    try:
+        decoded = base64.b64decode(m.group(1)).decode("utf-8")
+        return json.loads(decoded)
+    except Exception:  # noqa: BLE001
+        return None
+
+
+# ---------------------------------------------------------------------------
+# Prompt construction
+# ---------------------------------------------------------------------------
+
+def _build_prompt(failures: list[dict]) -> str:
+    trimmed = [
+        {
+            **f,
+            "failedTests": (f.get("failedTests") or [])[:MAX_TITLES],
+            "flakyTests":  (f.get("flakyTests")  or [])[:MAX_TITLES],
+        }
+        for f in failures
+    ]
+
+    return f"""You are analyzing Playwright test failures from a CI/CD pipeline.
+
+For each failing flow below, classify the root cause and provide actionable guidance.
+
+Failures:
+{json.dumps(trimmed, indent=2)}
+
+Respond with ONLY a raw JSON array (no markdown fences, no explanation).
+Each element must have exactly these fields:
+  "flow"           - (string) the flowName from the input
+  "job"            - (string) the job short name from the input
+  "classification" - one of: "service_regression", "environment_issue", "test_issue", "flaky"
+  "severity"       - one of: "high", "medium", "low"
+  "summary"        - (string <=120 chars) what is failing and likely why, inferred from test names
+  "recommendation" - (string <=200 chars) concrete next step to fix or investigate
+
+Classification guide:
+  service_regression  - feature tests failing; likely a code regression in the app
+  environment_issue   - infra/network/auth failures or env-specific issues
+  test_issue          - stale selectors, bad test data, or a test code bug
+  flaky               - intermittent with no clear pattern
+
+Start your response with [ and end with ].
+"""
+
+
+# ---------------------------------------------------------------------------
+# JSON extraction
+# ---------------------------------------------------------------------------
+
+def _extract_json(text: str) -> list | None:
+    """Robustly extract a JSON array from code-puppy output."""
+    osc = _extract_json_from_osc52(text)
+    if osc is not None:
+        return osc
+
+    stripped = _strip_ansi(text).strip()
+
+    if stripped.startswith("["):
+        try:
+            return json.loads(stripped)
+        except json.JSONDecodeError:
+            pass
+
+    m = re.search(r"```(?:json)?\s*(\[.*?\])\s*```", stripped, re.DOTALL)
+    if m:
+        try:
+            return json.loads(m.group(1))
+        except json.JSONDecodeError:
+            pass
+
+    m = re.search(r"(\[.*\])", stripped, re.DOTALL)
+    if m:
+        try:
+            return json.loads(m.group(1))
+        except json.JSONDecodeError:
+            pass
+
+    return None
+
+
+# ---------------------------------------------------------------------------
+# Public API
+# ---------------------------------------------------------------------------
+
+def analyze_failures(
+    failures: list[dict],
+    code_puppy_bin: str | None = None,
+) -> list[dict]:
+    """Invoke playwright-monocart-helper to classify each failing/flaky flow.
+
+    Args:
+        failures: List of failure dicts from save_failure_summary.
+        code_puppy_bin: Override path to code-puppy binary.
+
+    Returns:
+        List of analysis dicts, or [] on any failure (never raises).
+    """
+    if not failures:
+        return []
+
+    bin_path = Path(code_puppy_bin or _VENV_BIN)
+    if not bin_path.exists():
+        print(f"Warning: code-puppy not found at {bin_path}", file=sys.stderr)
+        return []
+
+    prompt = _build_prompt(failures)
+    n = len(failures)
+    print(f"Invoking {AGENT} to analyze {n} failing flow{'s' if n != 1 else ''}...", file=sys.stderr)
+
+    try:
+        result = subprocess.run(
+            [str(bin_path), "--prompt", prompt, "--agent", AGENT],
+            capture_output=True,
+            text=True,
+            timeout=TIMEOUT,
+        )
+    except subprocess.TimeoutExpired:
+        print(f"Warning: AI analysis timed out after {TIMEOUT}s.", file=sys.stderr)
+        return []
+    except Exception as exc:  # noqa: BLE001
+        print(f"Warning: AI analysis subprocess error: {exc}", file=sys.stderr)
+        return []
+
+    raw = result.stdout or ""
+    analysis = _extract_json(raw)
+
+    if analysis is None and result.stderr:
+        analysis = _extract_json(result.stderr)
+
+    if analysis is None:
+        print("Warning: Could not parse JSON from AI output.", file=sys.stderr)
+        tail = (raw or result.stderr or "").strip().splitlines()
+        for line in tail[-8:]:
+            print(f"   | {line}", file=sys.stderr)
+        return []
+
+    print(f"AI analysis complete — {len(analysis)} flow{'s' if len(analysis) != 1 else ''} classified.", file=sys.stderr)
+    return analysis
+
+
+# ---------------------------------------------------------------------------
+# CLI (standalone use)
+# ---------------------------------------------------------------------------
+
+def main() -> None:
+    import argparse
+
+    parser = argparse.ArgumentParser(
+        description="Analyze Playwright/Monocart failures with playwright-monocart-helper AI",
+    )
+    parser.add_argument("--failures", default="failures.json", help="Path to failures JSON")
+    parser.add_argument("--output", default="ai-analysis.json", help="Output path for analysis JSON")
+    parser.add_argument("--code-puppy-bin", default=None, help=f"Override path to code-puppy (default: {_VENV_BIN})")
+    args = parser.parse_args()
+
+    failures_path = Path(args.failures)
+    if not failures_path.exists():
+        print(f"Failures file not found: {failures_path}", file=sys.stderr)
+        sys.exit(1)
+
+    failures = json.loads(failures_path.read_text())
+    if not failures:
+        print("No failures to analyze.", file=sys.stderr)
+        sys.exit(0)
+
+    analysis = analyze_failures(failures, code_puppy_bin=args.code_puppy_bin)
+    if not analysis:
+        sys.exit(1)
+
+    output_path = Path(args.output)
+    output_path.write_text(json.dumps(analysis, indent=2))
+    print(f"Analysis saved to {output_path.resolve()}", file=sys.stderr)
+
+
+if __name__ == "__main__":
+    main()