resume-agent/backend/tests/test_resume_import_structure.py

"""Structured fallback and quality-gate coverage for resume imports."""

from __future__ import annotations

from typing import Any

from app.import_parser import ImportParseOutput, OpenAIResumeImportParser
from app.resume_import_service import RuleBasedResumeImportParser


class FakeCompletion:
    def __init__(self, result: ImportParseOutput) -> None:
        self.result = result
        self.calls: list[dict[str, Any]] = []

    def complete(self, **kwargs: Any) -> ImportParseOutput:
        self.calls.append(kwargs)
        return self.result


def _sample_resume() -> str:
    long_detail = "A" * 650
    return "\n".join(
        [
            "Li Ming",
            "li.ming@example.com | 13800138000 | Guangzhou",
            "Education",
            "Example University | Computer Science | Bachelor | 2022-09 - 2026-06",
            "GPA 3.8/4.0; ranked in the top 10%.",
            "Project Experience",
            "Resume Copilot | Backend Developer | 2025-01 - 2025-06",
            f"Built the resume parsing API and optimization workflow. {long_detail}",
            "\u5b9e\u4e60\u7ecf\u5386",
            "Example Tech | AI Engineering Intern | 2025-07 - 2025-09",
            "Implemented evaluation scripts and integrated retrieval.",
            "Skills",
            "Python, FastAPI, PostgreSQL, Docker",
        ]
    )


def test_rule_parser_structures_sections_and_does_not_truncate_text() -> None:
    draft = RuleBasedResumeImportParser().parse(
        source_name="resume.docx", text=_sample_resume()
    )

    document = draft.document
    assert document["basics"]["name"] == "Li Ming"
    assert document["basics"]["email"] == "li.ming@example.com"
    assert document["basics"]["phone"] == "13800138000"
    assert [section["kind"] for section in document["sections"]] == [
        "education",
        "project_experience",
        "internship_experience",
    ]
    project = document["sections"][1]["items"][0]
    assert project["project_name"] == "Resume Copilot"
    assert len(project["description"]) > 650
    assert document["skill_groups"]
    assert any("Python" in group["skills"] for group in document["skill_groups"])


def test_llm_skill_only_result_is_completed_with_local_sections() -> None:
    completion = FakeCompletion(
        ImportParseOutput.model_validate(
            {
                "basics": {},
                "target": {},
                "sections": [],
                "skill_groups": [
                    {"category": "\u7f16\u7a0b\u8bed\u8a00\u4e0e\u6846\u67b6", "skills": ["Python"]}
                ],
            }
        )
    )
    parser = OpenAIResumeImportParser(
        completion=completion,
        fallback=RuleBasedResumeImportParser(),
    )

    draft = parser.parse(source_name="resume.docx", text=_sample_resume())

    assert completion.calls
    assert {section["kind"] for section in draft.document["sections"]} >= {
        "education",
        "project_experience",
        "internship_experience",
    }
    assert draft.document["basics"]["phone"] == "13800138000"

def test_llm_unstructured_blob_is_replaced_by_detected_sections() -> None:
    completion = FakeCompletion(
        ImportParseOutput.model_validate(
            {
                "basics": {},
                "target": {},
                "sections": [
                    {
                        "kind": "additional_experience",
                        "heading": "导入内容",
                        "items": [{"fields": {"title": "resume.docx", "description": "raw text"}}],
                    }
                ],
                "skill_groups": [],
            }
        )
    )
    parser = OpenAIResumeImportParser(
        completion=completion,
        fallback=RuleBasedResumeImportParser(),
    )

    draft = parser.parse(source_name="resume.docx", text=_sample_resume())

    assert [section["kind"] for section in draft.document["sections"]] == [
        "education",
        "project_experience",
        "internship_experience",
    ]


def test_llm_backfill_preserves_each_project_and_original_summary() -> None:
    resume_text = "\n".join(
        [
            "Li Ming",
            "li.ming@example.com | 13800138000",
            "Project Experience",
            "Project Alpha | Backend Developer | 2025-01 - 2025-03",
            "Built the first service.",
            "Project Beta | Platform Engineer | 2025-04 - 2025-06",
            "Built the second service.",
            "Personal Summary",
            "Original summary paragraph one.",
            "Original summary paragraph two.",
        ]
    )
    completion = FakeCompletion(
        ImportParseOutput.model_validate(
            {
                "basics": {"name": "Li Ming"},
                "target": {},
                "profile_summary": "rewritten summary",
                "sections": [
                    {
                        "kind": "project_experience",
                        "heading": "Project Experience",
                        "items": [{"fields": {"project_name": "Project Alpha"}}],
                    }
                ],
                "skill_groups": [],
            }
        )
    )
    parser = OpenAIResumeImportParser(completion=completion, fallback=RuleBasedResumeImportParser())

    draft = parser.parse(source_name="resume.docx", text=resume_text)

    projects = next(section for section in draft.document["sections"] if section["kind"] == "project_experience")
    assert [item["project_name"] for item in projects["items"]] == ["Project Alpha", "Project Beta"]
    assert draft.document["profile_summary"] == {
        "content": "Original summary paragraph one.\nOriginal summary paragraph two.",
        "source": "user_edited",
        "generated_at": None,
        "stale": False,
    }


def test_llm_discards_unidentified_entries_and_merges_duplicate_education() -> None:
    resume_text = "\n".join(
        [
            "Li Ming",
            "li.ming@example.com | 13800138000",
            "Education",
            "Example University | Computer Science | Bachelor | 2022-09 - 2026-06",
            "GPA 3.8/4.0; ranked in the top 10%.",
            "Project Experience",
            "Project Alpha | Backend Developer | 2025-01 - 2025-03",
            "Built the first service.",
            "Project Beta | Platform Engineer | 2025-04 - 2025-06",
            "Built the second service.",
        ]
    )
    completion = FakeCompletion(
        ImportParseOutput.model_validate(
            {
                "basics": {"phone": "[redacted]", "email": "redacted@example.com"},
                "target": {},
                "sections": [
                    {
                        "kind": "education",
                        "heading": "Education",
                        "items": [
                            {"fields": {"school": "Example University"}},
                            {"fields": {"description": "orphaned education detail"}},
                        ],
                    },
                    {
                        "kind": "project_experience",
                        "heading": "Project Experience",
                        "items": [
                            {"fields": {"description": "orphaned project detail"}},
                            {"fields": {"project_name": "Project Alpha"}},
                        ],
                    },
                ],
                "skill_groups": [],
            }
        )
    )
    parser = OpenAIResumeImportParser(completion=completion, fallback=RuleBasedResumeImportParser())

    draft = parser.parse(source_name="resume.docx", text=resume_text)

    education = next(section for section in draft.document["sections"] if section["kind"] == "education")
    projects = next(section for section in draft.document["sections"] if section["kind"] == "project_experience")
    assert len(education["items"]) == 1
    assert education["items"][0]["school"] == "Example University"
    assert education["items"][0]["major"] == "Computer Science"
    assert "description" not in education["items"][0] or education["items"][0]["description"] != "orphaned education detail"
    assert [item["project_name"] for item in projects["items"]] == ["Project Alpha", "Project Beta"]
    assert draft.document["basics"]["phone"] == "13800138000"
    assert draft.document["basics"]["email"] == "li.ming@example.com"
    assert draft.document["import_metadata"]["parse_status"] == "needs_review"