fix: persist feedback model and ruleset versions#47
Conversation
There was a problem hiding this comment.
Pull request overview
Ensures scan feedback records include the associated model_version and ruleset_version so feedback can be audited and analyzed against the exact scan context (closes #44).
Changes:
- Add safe extraction/resolution helpers to pull
model_version/ruleset_versionfrom cached or persisted scan payloads before saving feedback. - Populate
ruleset_versioninto the stored scan “summary” payload and set a defaultmodel_versionin generated executive summaries. - Add API tests asserting feedback persistence of version fields and safe handling when metadata is missing.
Reviewed changes
Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
tests/api/test_feedback_endpoint.py |
Adds endpoint tests verifying save_feedback() receives extracted model_version/ruleset_version and handles missing metadata. |
src/extension_shield/core/summary_generator.py |
Adds model_version defaulting in the executive summary payload. |
src/extension_shield/api/main.py |
Adds version extraction helpers, sets ruleset_version on scan summary payloads, and persists versions on feedback save. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| ) | ||
|
|
||
|
|
||
| summary.setdefault("model_version", model_name) |
There was a problem hiding this comment.
summary.setdefault("model_version", model_name) may be misleading for auditability because invoke_with_fallback() can map the requested model_name to a provider-specific model name (and may fall back across providers). If you intend model_version to represent the actual model used, consider plumbing the resolved provider/model back from invoke_with_fallback (or otherwise recording it) rather than the requested env value.
| summary.setdefault("model_version", model_name) | |
| # Record the actual model used, if provided by the client response | |
| resolved_model = getattr(response, "model", None) or getattr(response, "model_name", None) | |
| if resolved_model: | |
| summary.setdefault("model_version", resolved_model) |
| ruleset_version="v1", | ||
| ) | ||
|
|
||
|
|
There was a problem hiding this comment.
The new feedback version behavior is only tested with payloads that have model_version in summary and scoring_v2 at the top level. In the SQLite backend, scoring_v2/report_view_model are commonly embedded under the summary JSON (and legacy rows may only have versions under summary["report_view_model"]["meta"]). Adding a test that exercises the DB payload shape (nested summary fields) would help ensure versions are persisted correctly across backends and older scan rows.
| def test_feedback_persists_versions_from_nested_summary_payload( | |
| client: TestClient, | |
| ) -> None: | |
| scan_id = "nestedabcdefghijnestedabcdefghij" | |
| scan_results.pop(scan_id, None) | |
| # Simulate a DB-style payload where scoring_v2 is embedded under summary | |
| scan_payload = { | |
| "summary": { | |
| "model_version": "gpt-4.1-mini", | |
| "scoring_v2": { | |
| "scoring_version": "2.1.0", | |
| "weights_version": "rules-v3", | |
| }, | |
| "report_view_model": { | |
| "meta": { | |
| "model_version": "legacy-model", | |
| "ruleset_version": "legacy-rules", | |
| } | |
| }, | |
| } | |
| } | |
| with patch("extension_shield.api.main.db") as mock_db: | |
| mock_db.get_scan_result = MagicMock(return_value=scan_payload) | |
| mock_db.save_feedback = MagicMock() | |
| response = client.post( | |
| "/api/feedback", | |
| json={ | |
| "scan_id": scan_id, | |
| "helpful": False, | |
| "reason": "score_off", | |
| "suggested_score": 85, | |
| "comment": "Nested summary payload", | |
| }, | |
| ) | |
| assert response.status_code == 200 | |
| mock_db.save_feedback.assert_called_once_with( | |
| scan_id=scan_id, | |
| helpful=False, | |
| reason="score_off", | |
| suggested_score=85, | |
| comment="Nested summary payload", | |
| user_id=None, | |
| model_version="gpt-4.1-mini", | |
| ruleset_version="rules-v3", | |
| ) |
|
|
||
| metadata = _coerce_json_dict(payload.get("metadata")) | ||
| summary = _coerce_json_dict(payload.get("summary")) | ||
| report_view_model = _coerce_json_dict(payload.get("report_view_model")) |
There was a problem hiding this comment.
_extract_feedback_versions() looks for report_view_model/meta only at the top level (payload["report_view_model"]). For the SQLite backend, report_view_model is persisted inside the summary JSON (see Database.save_scan_result storing report_view_model under summary), so report_meta will always be empty when feedback resolves via db.get_scan_result(). Consider also checking summary.get("report_view_model") (and then its meta) so version extraction works consistently across backends/legacy rows.
| report_view_model = _coerce_json_dict(payload.get("report_view_model")) | |
| report_view_model_source = payload.get("report_view_model") | |
| if not report_view_model_source: | |
| report_view_model_source = summary.get("report_view_model") | |
| report_view_model = _coerce_json_dict(report_view_model_source) |
|
@Th-Shivam can you make sure all the checks are passing and resubmit, thanks |
Stanzin7
added
help wanted
github-actions
Bot
added
the
area: backend
|
@Th-Shivam Any updates? |
|
I am really very sorry im in the middle of my term end examinations , i will solve these issues once my exams are over , most probably on 29th i will fix and raise the pr . |
|
@Th-Shivam That's completely fine; however, I don't think this needs much time. You can just run. npm audit --fix to get it fixed |
📝 WalkthroughWalkthroughAdds helpers to extract Changes
Sequence Diagram(s)sequenceDiagram
participant Client as Client
participant API as API (/api/feedback)
participant ScanStore as Scan Store (get_scan_result)
participant DB as DB (save_feedback)
Client->>API: POST /api/feedback {scan_id, feedback, user:"anon" / user_id}
API->>ScanStore: get_scan_result(scan_id)
ScanStore-->>API: scan_payload (summary, metadata, scoring_v2...)
API->>API: _coerce_json_fields() / _resolve_feedback_versions()
API->>DB: save_feedback({..., model_version, ruleset_version, user_id=None})
DB-->>API: saved
API-->>Client: 200 OK
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
github-actions
Bot
added
bug
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 3
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
src/extension_shield/core/summary_generator.py (1)
499-613:⚠️ Potential issue | 🟠 MajorDon't drop
model_versionon post-invoke failures.Once
invoke_with_fallback()succeeds, a parse/validation exception still falls through toreturn None. The scan payload then gets an empty summary, so feedback losesmodel_versioneven though it was already resolved from the response. Prefer returning a deterministic fallback summary with the attached version in this branch too.Suggested fix
- try: + resolved_model_version = None + try: # Format prompt to messages formatted_prompt = prompt.format_prompt() messages = formatted_prompt.to_messages() # Invoke with fallback response = invoke_with_fallback( messages=messages, model_name=model_name, model_parameters=model_parameters, ) resolved_model_version = _extract_response_model_version(response) @@ except Exception as exc: # Avoid noisy stack traces in normal operation; callers can decide how to handle None. logger.warning("Failed to generate executive summary: %s", exc) - return None + if facts_score is not None and facts_score_label is not None: + from extension_shield.core.report_view_model import _fallback_executive_summary + return _attach_model_version( + _fallback_executive_summary( + score=facts_score, + score_label=facts_score_label, + host_scope_label=facts_host_scope_label or "UNKNOWN", + ), + resolved_model_version, + ) + return {"model_version": resolved_model_version} if resolved_model_version else None🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/extension_shield/core/summary_generator.py` around lines 499 - 613, Initialize resolved_model_version = None before the try block, then after invoke_with_fallback keep assigning resolved_model_version as you already do; in the except Exception handler, if resolved_model_version is not None import and return the deterministic fallback via _attach_model_version(_fallback_executive_summary(...), resolved_model_version) (use the same score/score_label/host_scope_label logic you use earlier or pass defaults), otherwise keep the current logger.warning and return None; references: invoke_with_fallback, resolved_model_version, _attach_model_version, _fallback_executive_summary.
🧹 Nitpick comments (2)
tests/api/test_feedback_endpoint.py (1)
14-126: Add one cache-hit case forscan_resultstoo.These tests only cover the DB fallback, but
_resolve_feedback_versions()checks the in-memoryscan_resultscache first. A small happy-path case there would lock down the primary lookup path as well.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@tests/api/test_feedback_endpoint.py` around lines 14 - 126, Add a new unit test that exercises the in-memory cache path by populating the global scan_results dict for a test scan_id with a payload containing model_version and ruleset_version, then POST to /api/feedback and assert mock_db.save_feedback is called with those versions; ensure you reference the in-memory cache symbol scan_results and the feedback resolver (_resolve_feedback_versions) behavior, and optionally assert that mock_db.get_scan_result was not called to confirm the cache hit.src/extension_shield/api/main.py (1)
1566-1574: Reuse the normalized summary when buildingreport_view_model.This block adds
ruleset_versiontoexecutive_summary, but thereport_view_modelbuilt below still usesfinal_state.get("executive_summary"). That leaves the two payload shapes out of sync for the same scan.Suggested fix
- "report_view_model": build_report_view_model_safe( - manifest=manifest, - analysis_results={**analysis_results, "executive_summary": final_state.get("executive_summary") or {}}, + "report_view_model": build_report_view_model_safe( + manifest=manifest, + analysis_results={**analysis_results, "executive_summary": executive_summary}, metadata=metadata, extension_id=extension_id, scan_id=extension_id, ),🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/extension_shield/api/main.py` around lines 1566 - 1574, The executive_summary dict is normalized and updated with ruleset_version but the later report_view_model still reads final_state.get("executive_summary"), causing inconsistent payloads; update the code to reuse the normalized executive_summary when building report_view_model (either assign final_state["executive_summary"]=executive_summary after normalization or have report_view_model reference the local executive_summary variable) so both places use the same mutated object (refer to the executive_summary variable and the report_view_model construction).
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@src/extension_shield/api/main.py`:
- Around line 1036-1043: _resolve_feedback_versions currently only checks the
in-memory scan_results and db.get_scan_result, so file-backed scan payloads
loaded by get_scan_results(scan_id) are ignored; update
_resolve_feedback_versions to call get_scan_results(scan_id) as a final fallback
when payload is not found in scan_results and db.get_scan_result returns falsy,
then pass that payload into _extract_feedback_versions (handle None safely).
Refer to symbols: _resolve_feedback_versions, scan_results, db.get_scan_result,
get_scan_results, and _extract_feedback_versions.
In `@src/extension_shield/core/summary_generator.py`:
- Around line 64-69: The function _attach_model_version currently uses
normalized.setdefault("model_version", model_version) which preserves any
existing model_version in the parsed summary; instead, make the
transport/client-provided model_version authoritative by assigning it directly.
In _attach_model_version, keep the shallow-copy behavior (normalized =
dict(summary) if isinstance(summary, dict) else {}), then replace the setdefault
call with a direct assignment normalized["model_version"] = model_version when
model_version is truthy so the returned summary always reflects the resolved
model_version from the client.
In `@tests/api/test_feedback_endpoint.py`:
- Around line 39-47: The test assertions use the wrong anonymous user_id;
submit_feedback() calls _get_user_id() which returns "anon" for unauthenticated
requests, so update the mock_db.save_feedback assertions in these tests
(referencing the mock_db.save_feedback call in
tests/api/test_feedback_endpoint.py) to expect user_id="anon" (or alternatively
stub/authenticate _get_user_id in the test setup) so the asserted arguments
match submit_feedback() behavior.
---
Outside diff comments:
In `@src/extension_shield/core/summary_generator.py`:
- Around line 499-613: Initialize resolved_model_version = None before the try
block, then after invoke_with_fallback keep assigning resolved_model_version as
you already do; in the except Exception handler, if resolved_model_version is
not None import and return the deterministic fallback via
_attach_model_version(_fallback_executive_summary(...), resolved_model_version)
(use the same score/score_label/host_scope_label logic you use earlier or pass
defaults), otherwise keep the current logger.warning and return None;
references: invoke_with_fallback, resolved_model_version, _attach_model_version,
_fallback_executive_summary.
---
Nitpick comments:
In `@src/extension_shield/api/main.py`:
- Around line 1566-1574: The executive_summary dict is normalized and updated
with ruleset_version but the later report_view_model still reads
final_state.get("executive_summary"), causing inconsistent payloads; update the
code to reuse the normalized executive_summary when building report_view_model
(either assign final_state["executive_summary"]=executive_summary after
normalization or have report_view_model reference the local executive_summary
variable) so both places use the same mutated object (refer to the
executive_summary variable and the report_view_model construction).
In `@tests/api/test_feedback_endpoint.py`:
- Around line 14-126: Add a new unit test that exercises the in-memory cache
path by populating the global scan_results dict for a test scan_id with a
payload containing model_version and ruleset_version, then POST to /api/feedback
and assert mock_db.save_feedback is called with those versions; ensure you
reference the in-memory cache symbol scan_results and the feedback resolver
(_resolve_feedback_versions) behavior, and optionally assert that
mock_db.get_scan_result was not called to confirm the cache hit.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro Plus
Run ID: 870aa146-113a-4dac-90a5-e09589d5965c
⛔ Files ignored due to path filters (1)
frontend/package-lock.jsonis excluded by!**/package-lock.json
📒 Files selected for processing (3)
src/extension_shield/api/main.pysrc/extension_shield/core/summary_generator.pytests/api/test_feedback_endpoint.py
| def _resolve_feedback_versions(scan_id: str) -> tuple[Optional[str], Optional[str]]: | ||
| """Look up the related scan payload and extract version metadata safely.""" | ||
| payload = scan_results.get(scan_id) | ||
| if isinstance(payload, dict): | ||
| return _extract_feedback_versions(payload) | ||
|
|
||
| db_payload = db.get_scan_result(scan_id) | ||
| return _extract_feedback_versions(db_payload) |
There was a problem hiding this comment.
Include the existing file-backed scan fallback here too.
get_scan_results() can still load <scan_id>_results.json when the DB row is missing, but _resolve_feedback_versions() stops after memory/DB. That means feedback for file-only scans regresses to None again even though the metadata is still available.
Suggested fix
def _resolve_feedback_versions(scan_id: str) -> tuple[Optional[str], Optional[str]]:
"""Look up the related scan payload and extract version metadata safely."""
payload = scan_results.get(scan_id)
if isinstance(payload, dict):
return _extract_feedback_versions(payload)
- db_payload = db.get_scan_result(scan_id)
- return _extract_feedback_versions(db_payload)
+ try:
+ db_payload = db.get_scan_result(scan_id)
+ if isinstance(db_payload, dict):
+ return _extract_feedback_versions(db_payload)
+ except Exception as exc:
+ logger.debug("[FEEDBACK] Failed DB lookup for %s: %s", scan_id, exc)
+
+ result_file = RESULTS_DIR / f"{scan_id}_results.json"
+ if result_file.exists():
+ try:
+ with open(result_file, "r", encoding="utf-8") as f:
+ return _extract_feedback_versions(json.load(f))
+ except (OSError, json.JSONDecodeError) as exc:
+ logger.debug("[FEEDBACK] Failed file lookup for %s: %s", scan_id, exc)
+
+ return None, None🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@src/extension_shield/api/main.py` around lines 1036 - 1043,
_resolve_feedback_versions currently only checks the in-memory scan_results and
db.get_scan_result, so file-backed scan payloads loaded by
get_scan_results(scan_id) are ignored; update _resolve_feedback_versions to call
get_scan_results(scan_id) as a final fallback when payload is not found in
scan_results and db.get_scan_result returns falsy, then pass that payload into
_extract_feedback_versions (handle None safely). Refer to symbols:
_resolve_feedback_versions, scan_results, db.get_scan_result, get_scan_results,
and _extract_feedback_versions.
| mock_db.save_feedback.assert_called_once_with( | ||
| scan_id=scan_id, | ||
| helpful=False, | ||
| reason="score_off", | ||
| suggested_score=72, | ||
| comment="Score feels too strict", | ||
| user_id=None, | ||
| model_version="gpt-4o", | ||
| ruleset_version="v1", |
There was a problem hiding this comment.
These assertions expect the wrong anonymous user_id.
submit_feedback() calls _get_user_id(), and unauthenticated requests currently resolve to "anon", not None. Without stubbing auth or sending a user header, this assertion pattern will fail here and in the two tests below.
Suggested fix
- user_id=None,
+ user_id="anon",📝 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.
| mock_db.save_feedback.assert_called_once_with( | |
| scan_id=scan_id, | |
| helpful=False, | |
| reason="score_off", | |
| suggested_score=72, | |
| comment="Score feels too strict", | |
| user_id=None, | |
| model_version="gpt-4o", | |
| ruleset_version="v1", | |
| mock_db.save_feedback.assert_called_once_with( | |
| scan_id=scan_id, | |
| helpful=False, | |
| reason="score_off", | |
| suggested_score=72, | |
| comment="Score feels too strict", | |
| user_id="anon", | |
| model_version="gpt-4o", | |
| ruleset_version="v1", |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@tests/api/test_feedback_endpoint.py` around lines 39 - 47, The test
assertions use the wrong anonymous user_id; submit_feedback() calls
_get_user_id() which returns "anon" for unauthenticated requests, so update the
mock_db.save_feedback assertions in these tests (referencing the
mock_db.save_feedback call in tests/api/test_feedback_endpoint.py) to expect
user_id="anon" (or alternatively stub/authenticate _get_user_id in the test
setup) so the asserted arguments match submit_feedback() behavior.
There was a problem hiding this comment.
Actionable comments posted: 1
♻️ Duplicate comments (1)
src/extension_shield/api/main.py (1)
1036-1043:⚠️ Potential issue | 🟠 MajorKeep the file fallback and swallow lookup failures here.
Line 1042 still falls straight through to
db.get_scan_result(scan_id). If the scan only exists as<scan_id>_results.json, feedback will still persistNoneversions, and a DB lookup exception can now fail/api/feedbackoutright instead of degrading gracefully.Patch sketch
def _resolve_feedback_versions(scan_id: str) -> tuple[Optional[str], Optional[str]]: """Look up the related scan payload and extract version metadata safely.""" payload = scan_results.get(scan_id) if isinstance(payload, dict): return _extract_feedback_versions(payload) - db_payload = db.get_scan_result(scan_id) - return _extract_feedback_versions(db_payload) + try: + db_payload = db.get_scan_result(scan_id) + if isinstance(db_payload, dict): + return _extract_feedback_versions(db_payload) + except Exception as exc: + logger.debug("[FEEDBACK] Failed DB lookup for %s: %s", scan_id, exc) + + result_file = RESULTS_DIR / f"{scan_id}_results.json" + if result_file.exists(): + try: + with open(result_file, "r", encoding="utf-8") as f: + return _extract_feedback_versions(json.load(f)) + except (OSError, json.JSONDecodeError) as exc: + logger.debug("[FEEDBACK] Failed file lookup for %s: %s", scan_id, exc) + + return None, None🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/extension_shield/api/main.py` around lines 1036 - 1043, The current _resolve_feedback_versions falls through to db.get_scan_result(scan_id) without handling DB lookup failures or missing DB rows; wrap the DB call in a try/except and also treat a falsy db result as a fallback to the in-memory payload: keep the initial payload = scan_results.get(scan_id) check and return _extract_feedback_versions(payload) if it's a dict, otherwise try to get db_payload = db.get_scan_result(scan_id) inside a try block, and if that call raises or returns None/false, return _extract_feedback_versions(payload) (or _extract_feedback_versions(None) if payload is absent); reference symbols: _resolve_feedback_versions, scan_results, db.get_scan_result, and _extract_feedback_versions.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@src/extension_shield/api/main.py`:
- Around line 1566-1574: The current use of
executive_summary.setdefault("ruleset_version", ...) will leave an upstream None
or empty-string in place; change the logic in the block that builds
executive_summary so you explicitly check for a falsy value (e.g., if not
executive_summary.get("ruleset_version")) and then assign
executive_summary["ruleset_version"] = scoring_v2_payload.get("weights_version")
or scoring_v2_payload.get("scoring_version"). Update the code around the
executive_summary creation (the executive_summary variable and the setdefault
call) to use this truthy check so a None/"" is overwritten by the resolved
scoring_v2_payload value.
---
Duplicate comments:
In `@src/extension_shield/api/main.py`:
- Around line 1036-1043: The current _resolve_feedback_versions falls through to
db.get_scan_result(scan_id) without handling DB lookup failures or missing DB
rows; wrap the DB call in a try/except and also treat a falsy db result as a
fallback to the in-memory payload: keep the initial payload =
scan_results.get(scan_id) check and return _extract_feedback_versions(payload)
if it's a dict, otherwise try to get db_payload = db.get_scan_result(scan_id)
inside a try block, and if that call raises or returns None/false, return
_extract_feedback_versions(payload) (or _extract_feedback_versions(None) if
payload is absent); reference symbols: _resolve_feedback_versions, scan_results,
db.get_scan_result, and _extract_feedback_versions.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro Plus
Run ID: 521f30e9-76ae-437d-9e56-b6394b281bc8
📒 Files selected for processing (3)
src/extension_shield/api/main.pysrc/extension_shield/core/summary_generator.pytests/api/test_feedback_endpoint.py
🚧 Files skipped from review as they are similar to previous changes (2)
- src/extension_shield/core/summary_generator.py
- tests/api/test_feedback_endpoint.py
| executive_summary = final_state.get("executive_summary") or {} | ||
| if isinstance(executive_summary, dict): | ||
| executive_summary = dict(executive_summary) | ||
| else: | ||
| executive_summary = {} | ||
| executive_summary.setdefault( | ||
| "ruleset_version", | ||
| scoring_v2_payload.get("weights_version") or scoring_v2_payload.get("scoring_version"), | ||
| ) |
There was a problem hiding this comment.
setdefault() leaves null/blank ruleset_version untouched.
If upstream already includes ruleset_version: None or "", Line 1571 treats that as present and persists the empty value even though scoring_v2_payload already resolved the version. A truthy check is safer here.
Patch sketch
- executive_summary.setdefault(
- "ruleset_version",
- scoring_v2_payload.get("weights_version") or scoring_v2_payload.get("scoring_version"),
- )
+ resolved_ruleset_version = (
+ scoring_v2_payload.get("weights_version")
+ or scoring_v2_payload.get("scoring_version")
+ )
+ if resolved_ruleset_version and not executive_summary.get("ruleset_version"):
+ executive_summary["ruleset_version"] = resolved_ruleset_version🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@src/extension_shield/api/main.py` around lines 1566 - 1574, The current use
of executive_summary.setdefault("ruleset_version", ...) will leave an upstream
None or empty-string in place; change the logic in the block that builds
executive_summary so you explicitly check for a falsy value (e.g., if not
executive_summary.get("ruleset_version")) and then assign
executive_summary["ruleset_version"] = scoring_v2_payload.get("weights_version")
or scoring_v2_payload.get("scoring_version"). Update the code around the
executive_summary creation (the executive_summary variable and the setdefault
call) to use this truthy check so a None/"" is overwritten by the resolved
scoring_v2_payload value.
|
@Th-Shivam Thanks for taking the initiative, I have fixed this on your behalf |
Closes #44
Summary by CodeRabbit
Bug Fixes
Improvements
Tests