[Fix] 자소서 분석 개선 예시 문장 생성 방식 수정#81
Merged
Merged
Conversation
- LLM 프롬프트에서 improvement를 완성된 개선 예시 문장으로 생성하도록 지시 강화 - improvement에 첨삭 조언이나 지시문이 들어오지 않도록 금지 규칙 추가 - 지시문 형태의 improvement가 반환되면 빈 값으로 정규화하도록 방어 로직 추가 - improvement 정규화 테스트 추가
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (1)
📝 WalkthroughWalkthroughThe PR tightens the AI prompt to require fully formed Korean replacement sentences for per-question "improvement" entries, adds service-side normalization to blank instruction-like LLM outputs, and extends tests to assert both blanking and preservation behaviors. ChangesImprovement field normalization
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Comment |
![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
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In
`@src/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisAiClient.java`:
- Around line 137-141: Update the prompt text in AnalysisAiClient to exactly
mirror the service-layer validation in
AnalysisService.isInstructionLikeImprovement(): include both verb forms like
"추가하세요" and "추가하여" (and other variants the service checks), and list the
sentence-ending patterns the service rejects (e.g., "하세요.", "해주세요.", "해야 합니다.")
so the LLM guidance and the validation logic are consistent; reference the
method AnalysisService.isInstructionLikeImprovement() when locating the exact
phrase/regex patterns to sync and modify the bullet points in the prompt block
accordingly.
In
`@src/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisService.java`:
- Around line 234-242: The isInstructionLikeImprovement method currently uses
brittle contains() checks that cause false positives; replace those substring
checks with stricter pattern matching (e.g., use a regex to detect
imperative/instructional sentence forms at boundaries or combine imperative
verbs with imperative endings) or invert to a whitelist of declarative endings;
specifically, modify isInstructionLikeImprovement to stop using
improvement.contains("추가하여") / "명확히 하세요" / "보완하세요" and instead use a regex that
matches instruction-like forms (imperative phrases ending with "하세요", "해주세요",
"하세요\\." etc. anchored to sentence boundaries) or validate declarative endings
such as "습니다.", "했습니다.", "됩니다." to avoid discarding valid declarative
improvements referenced by AnalysisAiClient prompt.
In
`@src/test/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisServiceTest.java`:
- Around line 239-264: Add the new test method
analyzePreservesDeclarativeImprovement (as in the suggested snippet) to
AnalysisServiceTest to assert that a legitimate declarative improvement string
is preserved; then update the isInstructionLikeImprovement(...) logic (the
method currently using contains("추가하여") and similar checks) to be more
discriminating—e.g., detect imperative forms or explicit instruction markers
(use regex that matches sentences starting with imperative verbs or ending with
imperative/formal instruction endings like "하세요", "해라", "해주세요") instead of
blanket contains checks—so declarative sentences such as the provided
validImprovement are not normalized to empty.
🪄 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: 5c97abe4-5b0f-49ae-9faa-60959cccaf24
📒 Files selected for processing (3)
src/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisAiClient.javasrc/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisService.javasrc/test/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisServiceTest.java
Comment on lines
+239
to
+264
| @Test | ||
| @DisplayName("LLM improvement가 첨삭 지시문이면 빈 값으로 저장한다") | ||
| void analyzeNormalizesInstructionLikeImprovement() { | ||
| User user = saveUser("analysis-instruction-improvement@example.com"); | ||
| MockApply mockApply = saveMockApply(user); | ||
| Question question = saveQuestion(mockApply, "문제 해결 경험", "저는 로그를 확인하고 쿼리 실행 계획을 분석했습니다."); | ||
| when(analysisAiClient.analyze(any(), any())).thenReturn(new AnalysisLlmResponse( | ||
| 64, | ||
| 70, | ||
| 55, | ||
| 67, | ||
| "개선 예시 문장 검증입니다.", | ||
| List.of(new AnalysisLlmResponse.QuestionAnalysisItem( | ||
| question.getId(), | ||
| "저는 로그를 확인하고 쿼리 실행 계획을 분석했습니다.", | ||
| "mentioned", | ||
| "성과 수치가 부족합니다.", | ||
| "성과 수치를 추가하여 문제 해결의 효과를 명확히 하세요." | ||
| )) | ||
| )); | ||
|
|
||
| AnalysisResponse response = analysisService.analyze(user, mockApply.getId()); | ||
|
|
||
| assertThat(response.questions().get(0).analyses()).hasSize(1); | ||
| assertThat(response.questions().get(0).analyses().get(0).improvement()).isEmpty(); | ||
| } |
There was a problem hiding this comment.
🛠️ Refactor suggestion | 🟠 Major | ⚡ Quick win
Expand test coverage to include false positive scenarios.
The current test validates that instruction-like improvements are correctly filtered, but doesn't verify that legitimate declarative sentences are preserved. Given the contains() checks in isInstructionLikeImprovement(), you should add test cases to ensure valid improvement sentences aren't incorrectly normalized to empty strings.
🧪 Suggested additional test case
`@Test`
`@DisplayName`("완성된 평서문 improvement는 그대로 저장한다")
void analyzePreservesDeclarativeImprovement() {
User user = saveUser("analysis-declarative-improvement@example.com");
MockApply mockApply = saveMockApply(user);
Question question = saveQuestion(mockApply, "문제 해결 경험", "저는 로그를 확인했습니다.");
String validImprovement = "저는 로그를 분석하고 누락된 인덱스를 추가하여 응답 시간을 1.8초에서 0.6초로 단축했습니다.";
when(analysisAiClient.analyze(any(), any())).thenReturn(new AnalysisLlmResponse(
70,
75,
65,
70,
"평서문 검증입니다.",
List.of(new AnalysisLlmResponse.QuestionAnalysisItem(
question.getId(),
"저는 로그를 확인했습니다.",
"mentioned",
"성과가 부족합니다.",
validImprovement
))
));
AnalysisResponse response = analysisService.analyze(user, mockApply.getId());
assertThat(response.questions().get(0).analyses()).hasSize(1);
assertThat(response.questions().get(0).analyses().get(0).improvement())
.isEqualTo(validImprovement);
}This test would currently fail due to the overly broad contains("추가하여") check.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In
`@src/test/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisServiceTest.java`
around lines 239 - 264, Add the new test method
analyzePreservesDeclarativeImprovement (as in the suggested snippet) to
AnalysisServiceTest to assert that a legitimate declarative improvement string
is preserved; then update the isInstructionLikeImprovement(...) logic (the
method currently using contains("추가하여") and similar checks) to be more
discriminating—e.g., detect imperative forms or explicit instruction markers
(use regex that matches sentences starting with imperative verbs or ending with
imperative/formal instruction endings like "하세요", "해라", "해주세요") instead of
blanket contains checks—so declarative sentences such as the provided
validImprovement are not normalized to empty.
- improvement 지시문 판별 로직에서 과도한 contains 조건 제거 - 명령형/지시문 어미 중심으로 improvement 필터링 기준 수정 - 정상적인 평서문 개선 예시가 빈 값으로 정규화되지 않도록 보완 - LLM 프롬프트의 improvement 금지 표현을 서버 검증 기준과 일치하도록 수정 - 완성된 평서문 improvement 보존 테스트 추가
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
src/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisService.java (1)
235-241:⚠️ Potential issue | 🟡 Minor | ⚡ Quick winCover spaced imperative form (
해 주세요) in normalization.Current checks catch
해주세요.but can miss해 주세요.variants, allowing instruction-like text through.Suggested patch
private boolean isInstructionLikeImprovement(String improvement) { return improvement.endsWith("하세요.") || improvement.endsWith("해주세요.") || improvement.endsWith("해야 합니다.") || improvement.endsWith("필요합니다.") + || improvement.matches(".*해\\s*주세요\\.?$") || improvement.matches(".*[을를]\\s+(추가|보완|수정)하(세요|십시오).*") || improvement.matches(".*명확히\\s+(하|해)(세요|야 합니다).*") - || improvement.matches("^(추가|보완|수정).*(하세요|해주세요)\\.$"); + || improvement.matches("^(추가|보완|수정).*(하세요|해\\s*주세요|해주세요)\\.$"); }🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisService.java` around lines 235 - 241, The current imperative-detection expression in AnalysisService (the return statement that checks improvement.endsWith(...) and improvement.matches(...)) misses spaced variants like "해 주세요" or "해 주십시오"; update those checks to accept optional whitespace between "하" and the following honorific/ending. Specifically, add endsWith("해 주세요.") and endsWith("해 주십시오.") (or use a single endsWith check with normalized whitespace), and change the regex fragments "하(세요|십시오)" and "하(세요|야 합니다)" to allow optional spaces (e.g., "하\\s*(세요|십시오)" and "하\\s*(세요|야 합니다)"), and similarly update any other matches that assume no space so the method in AnalysisService correctly flags both joined and spaced imperative forms.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Outside diff comments:
In
`@src/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisService.java`:
- Around line 235-241: The current imperative-detection expression in
AnalysisService (the return statement that checks improvement.endsWith(...) and
improvement.matches(...)) misses spaced variants like "해 주세요" or "해 주십시오";
update those checks to accept optional whitespace between "하" and the following
honorific/ending. Specifically, add endsWith("해 주세요.") and endsWith("해 주십시오.")
(or use a single endsWith check with normalized whitespace), and change the
regex fragments "하(세요|십시오)" and "하(세요|야 합니다)" to allow optional spaces (e.g.,
"하\\s*(세요|십시오)" and "하\\s*(세요|야 합니다)"), and similarly update any other matches
that assume no space so the method in AnalysisService correctly flags both
joined and spaced imperative forms.
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro Plus
Run ID: 592478b2-7dff-4d37-b500-3540ea73599c
📒 Files selected for processing (3)
src/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisAiClient.javasrc/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisService.javasrc/test/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisServiceTest.java
- improvement 지시문 감지 로직에 띄어쓰기 포함 표현 추가 - 해 주세요, 해 주십시오 형태의 지시문 어미를 필터링하도록 수정 - 명령형 정규식에서 하세요/하십시오 앞 공백을 허용하도록 보완 - 띄어쓰기 포함 첨삭 지시문 정규화 테스트 추가
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In
`@src/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisService.java`:
- Around line 237-242: The current normalization/filtering logic in
AnalysisService.java for the variable improvement misses imperative variants
ending with "하십시오" and "해주십시오", so phrases like "문장을 명확히 하십시오." or "수정해주십시오."
bypass the filter; update the conditional that checks improvement (the branch
with improvement.endsWith(...) and improvement.matches(...)) to include explicit
checks for "하십시오" and "해주십시오" variants and extend the regex patterns (e.g., the
matches that handle "해 주십시오", "해주십시오", and "하십시오" forms) so those imperative
forms are normalized/filtered the same as existing "해 주세요"/"해 주십시오"/"해야 합니다"
cases.
🪄 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: d5b1bdfe-dad2-49bc-a1fa-01e19d1132f9
📒 Files selected for processing (2)
src/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisService.javasrc/test/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisServiceTest.java
- 하십시오, 해주십시오 형태의 격식체 지시문 필터링 추가 - 명확히 하십시오 등 누락된 명령형 표현 정규식 보완 - 격식체 첨삭 지시문이 빈 improvement로 저장되는 회귀 테스트 추가
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In
`@src/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisService.java`:
- Line 245: The current improvement.matches(...) call in AnalysisService.java
forces a trailing period via "\\.$", causing imperative commands without
punctuation (e.g., "수정해주십시오") to bypass the filter; update the regex used in the
improvement.matches(...) expression so the final period is optional (for example
by using "(\\.|$)" or making the final "\\." optional) so both forms with and
without a trailing dot match; adjust only the regex string in the
improvement.matches(...) invocation to enforce this change.
🪄 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: e394f600-d11e-4e48-afcc-4280c80ea119
📒 Files selected for processing (2)
src/main/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisService.javasrc/test/java/com/jobdri/jobdri_api/domain/analysis/service/AnalysisServiceTest.java
- improvement 지시문 정규식의 마침표 조건을 선택 처리 - 수정해주십시오 / 수정해주십시오. 형태 모두 필터링되도록 수정
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
✨ 어떤 이유로 PR를 하셨나요?
📋 세부 내용 - 왜 해당 PR이 필요한지 작업 내용을 자세하게 설명해주세요
📸 작업 화면 스크린샷
🚨 관련 이슈 번호 [#24 ]
Summary by CodeRabbit