Add CSV export support for SuiteResultAdd csv reports

[Jagriti-student]

Summary

This PR adds a to_csv() method to SuiteResult in src/agentunit/reporting/results.py.
It allows exporting all test results to a CSV file, flattening nested metrics into separate columns
(e.g., metric_ExactMatch, metric_Latency) for easy analysis in spreadsheets.

Changes

• Added _flatten_metrics() helper (already existing but improved if needed)
• Added SuiteResult.to_csv(path: str | Path) method
• CSV writes one row per scenario run

Testing

Verified manually using:

from datetime import datetime
from agentunit.reporting.results import SuiteResult, ScenarioResult, ScenarioRun

run = ScenarioRun(
    scenario_name="DemoScenario",
    case_id="case_1",
    success=True,
    metrics={"ExactMatch": 1.0, "Latency": 120.5},
    duration_ms=150.0,
    trace=None,
)

scenario = ScenarioResult(name="DemoScenario")
scenario.add_run(run)

suite = SuiteResult(
    scenarios=[scenario],
    started_at=datetime.now(),
    finished_at=datetime.now(),
)

suite.to_csv("results/test.csv")

Closes #58  

<!-- This is an auto-generated comment: release notes by coderabbit.ai -->
## Summary by CodeRabbit

* **New Features**
  * CSV export for test results, including scenario, case ID, success, duration, error, and flattened metrics.

* **Improvements**
  * Reorganized JUnit output and improved time reporting accuracy.
  * Tighter Markdown and JSON formatting for exported reports.

* **Maintenance**
  * Removed numpy dependency and added an "integration-tests" extras entry; minor dependency reordering and formatting cleanup.

<sub>✏️ Tip: You can customize this high-level summary in your review settings.</sub>
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

[continue]

Learn more

---

All Green is an AI agent that automatically:

✅ Addresses code review comments

✅ Fixes failing CI checks

✅ Resolves merge conflicts

---

Unsubscribe from All Green comments

[coderabbitai]

Warning

Rate limit exceeded

@Jagriti-student has exceeded the limit for the number of commits that can be reviewed per hour. Please wait 14 minutes and 16 seconds before requesting another review.

⌛ How to resolve this issue?

After the wait time has elapsed, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

We recommend that you space out your commits to avoid hitting the rate limit.

🚦 How do rate limits work?

CodeRabbit enforces hourly rate limits for each developer per organization.

Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout.

Please see our FAQ for further information.

📥 Commits

Reviewing files that changed from the base of the PR and between fc43108 and 637bb2d.

📒 Files selected for processing (1)

• src/agentunit/reporting/results.py

Walkthrough

Adds CSV export to SuiteResult with metric flattening, refactors JUnit output to a single testsuite and adjusts time reporting, enhances Markdown rendering helpers, and updates pyproject.toml (removes numpy, reorders dependencies, adds integration-tests extra).

Changes

Cohort / File(s)	Summary
Configuration pyproject.toml	Removed numpy dependency, removed inline comment from langchain line, reordered jsonschema within dependencies, and added integration-tests = ["langgraph"] under [tool.poetry.extras].
Reporting / Exports src/agentunit/reporting/results.py	Added `SuiteResult.to_csv(path: str
Tests / Minor edits tests/test_reporting.py	Minor import/comment and whitespace cleanup in test file.

Sequence Diagram(s)

(omitted — changes are internal export additions and formatting; no multi-component sequential flow requiring visualization)

Estimated code review effort

🎯 4 (Complex) | ⏱️ ~45 minutes

Possibly related PRs

• Add LangGraph integration tests (#24) #36: Overlaps with pyproject.toml edits adding an integration-tests extra that includes langgraph.
• feat: add HTML report exporter for SuiteResult #48: Related changes to SuiteResult export code and HTML/report rendering.
• Feature/html report exporter #50: Related expansions to export functionality in src/agentunit/reporting/results.py.

Suggested reviewers

• aviralgarg05

🚥 Pre-merge checks | ✅ 2 | ❌ 3

❌ Failed checks (2 warnings, 1 inconclusive)

Check name	Status	Explanation	Resolution
Out of Scope Changes check	⚠️ Warning	Changes to pyproject.toml (numpy removal, jsonschema reorder, langchain comment removal, integration-tests extras) and test file formatting are out of scope relative to issue #58's CSV export objective.	Remove the pyproject.toml changes (unrelated dependency management) and test file formatting changes to keep the PR focused on CSV export functionality per issue #58.
Docstring Coverage	⚠️ Warning	Docstring coverage is 22.22% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.
Description check	❓ Inconclusive	The PR description includes a summary, changes, and testing verification but lacks the detailed template structure with Type of Change, Testing checklist, Code Quality checks, and other required sections.	Provide a more complete description using the template: include Type of Change, Testing checklist, Code Quality verification, and Documentation updates to align with repository standards.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly addresses the main change (CSV export) but contains awkward phrasing with duplication ('Add CSV export support for SuiteResultAdd csv reports').
Linked Issues check	✅ Passed	The PR successfully implements all requirements from issue #58: adds to_csv() method, flattens nested metrics, and writes one row per run with proper CSV structure.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

❌ Patch coverage is 11.76471% with 30 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/agentunit/reporting/results.py	11.76%	30 Missing ⚠️

📢 Thoughts on this report? Let us know!

[coderabbitai]

Actionable comments posted: 1

🧹 Nitpick comments (2)

src/agentunit/reporting/results.py (2)

83-95: One-level flattening is intentional but should be documented.

The helper only flattens one level of nesting. If metrics contains deeper structures like {"outer": {"inner": {"deep": 1}}}, the inner dict becomes a cell value (e.g., "{'deep': 1}"), which may not be spreadsheet-friendly.

If deeper nesting is a realistic scenario, consider making this recursive or adding a docstring clarifying the depth limit.

---

217-219: Column order is alphabetical — consider prioritizing core columns.

Sorting fieldnames alphabetically places case_id before scenario_name and interleaves metric_* columns unpredictably. For spreadsheet usability, leading with fixed columns (scenario_name, case_id, success, duration_ms, error) followed by sorted metric columns is more intuitive.

Proposed fix

-        fieldnames = sorted(
-            {key for row in rows for key in row.keys()}
-        )
+        base_fields = ["scenario_name", "case_id", "success", "duration_ms", "error"]
+        metric_fields = sorted(
+            {key for row in rows for key in row.keys()} - set(base_fields)
+        )
+        fieldnames = base_fields + metric_fields

📜 Review details

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 40bc42d and 580323d.

⛔ Files ignored due to path filters (1)

• poetry.lock is excluded by !**/*.lock

📒 Files selected for processing (3)

• pyproject.toml
• src/agentunit/reporting/results.py
• tests/test_reporting.py

💤 Files with no reviewable changes (1)

• tests/test_reporting.py

🧰 Additional context used

🧬 Code graph analysis (1)

src/agentunit/reporting/results.py (3)

src/agentunit/reporting/html.py (1)

• render_html_report (11-102)

src/agentunit/core/runner.py (1)

• run (45-54)

src/agentunit/datasets/base.py (1)

• name (38-39)

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (2)

• GitHub Check: Test (Python 3.12)
• GitHub Check: Test (Python 3.10)

🔇 Additional comments (4)

pyproject.toml (1)

39-41: LGTM — extras group aligns with test markers.

The new integration-tests extras entry correctly groups langgraph for optional testing scenarios, matching the pytest marker defined at line 70.

src/agentunit/reporting/results.py (3)

133-177: LGTM — JUnit structure is now a single testsuite element.

The refactored XML uses a single <testsuite> with nested <testcase> elements, which is a more standard JUnit format. Time reporting and failure message handling are implemented correctly.

---

189-226: CSV export implementation looks solid overall.

The method correctly:

• Creates parent directories
• Flattens metrics for spreadsheet compatibility
• Uses DictWriter with proper encoding and newline handling

Minor improvements noted in separate comments regarding column ordering and empty-file behavior.

---

268-272: No changes needed—the type annotation is correct.

ScenarioRun.metrics is properly typed as dict[str, float | None], which is consistent with the .2f format used in the markdown rendering. The _flatten_metrics function handles nested dicts, but it's used only in CSV export (line 210), not in the markdown rendering at lines 268-272. The markdown code correctly assumes float or None values.

Likely an incorrect or invalid review comment.

[coderabbitai]

⚠️ Potential issue | 🟡 Minor

Inconsistent behavior: empty suite doesn't create file.

When there are no runs, this returns early without writing any file. Other export methods (to_json, to_markdown, to_junit) always create the output file even when empty. This inconsistency could surprise callers who expect the file to exist after a successful call.

Proposed fix: create empty CSV with headers or at minimum an empty file

         if not rows:
+            target.touch()
             return target

Or, to maintain header consistency for empty results, define a fixed set of base fieldnames.

📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change

	if not rows:
	return target
	if not rows:
	target.touch()
	return target

[coderabbitai]

Actionable comments posted: 1

🤖 Fix all issues with AI agents

In @src/agentunit/reporting/results.py:
- Line 206: The comprehension building fieldnames uses row.keys (a method
object) instead of calling it; update the expression to call row.keys() so it
reads: fieldnames = sorted({key for row in rows for key in row.keys()}) to avoid
TypeError and correctly collect keys from each row in rows.

🧹 Nitpick comments (1)

src/agentunit/reporting/results.py (1)

203-204: Consider documenting the empty-suite behavior.

When there are no runs, the method returns without creating a file. This is reasonable since fieldnames are derived from row data, but callers might expect a file to exist. Consider documenting this in the docstring.

📝 Suggested docstring update

     def to_csv(self, path: str | Path) -> Path:
         """
         Export suite results to CSV.
         One row per scenario run.
+
+        Note: If there are no scenario runs, no file is created.
         """

📜 Review details

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 580323d and fc43108.

📒 Files selected for processing (1)

• src/agentunit/reporting/results.py

🧰 Additional context used

🧬 Code graph analysis (1)

src/agentunit/reporting/results.py (1)

src/agentunit/core/runner.py (1)

• run (45-54)

🔇 Additional comments (5)

src/agentunit/reporting/results.py (5)

5-5: LGTM!

The new imports (csv module and Any type) are appropriate for the CSV export functionality being added.

Also applies to: 10-10

---

82-92: Implementation handles single-level nesting only.

The helper correctly flattens one level of nested dictionaries. If metrics contain deeper nesting (e.g., {"a": {"b": {"c": 1}}}), inner dicts would be serialized as strings rather than flattened. This is likely sufficient for current use cases where ScenarioRun.metrics is typed as dict[str, float | None].

---

116-119: LGTM!

Explicit encoding="utf-8" is good practice for cross-platform consistency.

---

131-166: LGTM!

The simplified JUnit output with a single testsuite root element is a valid format widely supported by CI tools.

---

241-261: LGTM!

The explicit type hints and formatting improvements enhance readability without changing behavior.

[aviralgarg05]

LGTM!