job-research/tests/test_cli.py

from subprocess import run
from textwrap import dedent

from typer.testing import CliRunner

from job_research.cli import app
from job_research.storage import load_yaml


def test_installed_cli_help_exposes_build_profile_subcommand() -> None:
    result = run(["uv", "run", "job-research", "--help"], capture_output=True, text=True, check=False)

    assert result.returncode == 0
    assert "build-profile" in result.stdout
    assert "Build candidate-profile.yaml from CV and markdown profile." in result.stdout


def test_installed_cli_subcommand_help_works() -> None:
    result = run(["uv", "run", "job-research", "build-profile", "--help"], capture_output=True, text=True, check=False)

    assert result.returncode == 0
    assert "Usage: job-research build-profile" in result.stdout


def test_build_profile_writes_yaml_from_cv_and_profile(tmp_path) -> None:
    cv = tmp_path / "cv.txt"
    cv.write_text(
        dedent(
            """
            Tonio Example
            Location: France
            Languages: French, English
            Skills: Python, SQL, Terraform
            Years of experience: 3
            Data Engineer at Acme
            Education: Master of Science at Example University
            """
        ).strip(),
        encoding="utf-8",
    )
    profile = tmp_path / "profile.md"
    profile.write_text(
        dedent(
            """
            # Candidate Profile

            ## Summary
            Junior data engineer focused on Python and GCP.

            ## Target Roles
            - Data Engineer

            ## Strengths
            - Python
            - SQL

            ## Skills To Emphasize
            - GCP
            - BigQuery

            ## Constraints
            - CDI only
            - France only

            ## Notes
            - Slight preference for French listings.
            - Years of experience feels closer to 4.
            """
        ).strip(),
        encoding="utf-8",
    )
    out = tmp_path / "candidate-profile.yaml"

    result = run(
        ["uv", "run", "job-research", "build-profile", "--cv", str(cv), "--profile", str(profile), "--out", str(out)],
        capture_output=True,
        text=True,
        check=False,
    )

    assert result.returncode == 0
    assert f"candidate profile written to {out}" in result.stdout
    assert "Warnings included: 1" in result.stdout

    payload = load_yaml(out)

    assert payload["name"] == "Tonio Example"
    assert payload["summary"] == "Junior data engineer focused on Python and GCP."
    assert payload["target_roles"] == ["Data Engineer"]
    assert payload["skills"] == ["Python", "SQL", "Terraform", "GCP", "BigQuery"]
    assert any(item["field"] == "years_of_experience" for item in payload["warnings"])


def test_build_profile_reports_when_no_warnings_are_included(tmp_path) -> None:
    cv = tmp_path / "cv.txt"
    cv.write_text(
        dedent(
            """
            Tonio Example
            Location: France
            Languages: French, English
            Skills: Python, SQL
            Data Engineer at Acme
            Education: Master of Science at Example University
            """
        ).strip(),
        encoding="utf-8",
    )
    profile = tmp_path / "profile.md"
    profile.write_text(
        dedent(
            """
            # Candidate Profile

            ## Summary
            Junior data engineer focused on Python and GCP.

            ## Target Roles
            - Data Engineer

            ## Strengths
            - Python
            - SQL

            ## Skills To Emphasize
            - GCP
            - BigQuery

            ## Constraints
            - CDI only
            - France only

            ## Notes
            - Slight preference for French listings.
            """
        ).strip(),
        encoding="utf-8",
    )
    out = tmp_path / "candidate-profile.yaml"

    result = run(
        ["uv", "run", "job-research", "build-profile", "--cv", str(cv), "--profile", str(profile), "--out", str(out)],
        capture_output=True,
        text=True,
        check=False,
    )

    assert result.returncode == 0
    assert f"candidate profile written to {out}" in result.stdout
    assert "No warnings included." in result.stdout


def test_build_profile_rejects_empty_cv_text_before_writing(tmp_path) -> None:
    cv = tmp_path / "cv.txt"
    cv.write_text("   \n", encoding="utf-8")
    profile = tmp_path / "profile.md"
    profile.write_text(
        dedent(
            """
            # Candidate Profile

            ## Summary
            Junior data engineer.

            ## Target Roles
            - Data Engineer

            ## Strengths
            - Python

            ## Skills To Emphasize
            - BigQuery

            ## Constraints
            - CDI only

            ## Notes
            - Slight preference for French listings.
            """
        ).strip(),
        encoding="utf-8",
    )
    out = tmp_path / "candidate-profile.yaml"

    result = run(
        ["uv", "run", "job-research", "build-profile", "--cv", str(cv), "--profile", str(profile), "--out", str(out)],
        capture_output=True,
        text=True,
        check=False,
    )

    assert result.returncode != 0
    assert not out.exists()
    assert "No readable text found in CV input" in result.stderr
    assert "Traceback" not in result.stderr


def test_build_profile_reports_unreadable_pdf_input_cleanly(tmp_path, monkeypatch) -> None:
    cv = tmp_path / "cv.pdf"
    cv.write_bytes(b"%PDF-1.4\n")
    profile = tmp_path / "profile.md"
    profile.write_text(
        dedent(
            """
            # Candidate Profile

            ## Summary
            Junior data engineer.

            ## Target Roles
            - Data Engineer

            ## Strengths
            - Python

            ## Skills To Emphasize
            - BigQuery

            ## Constraints
            - CDI only

            ## Notes
            - Slight preference for French listings.
            """
        ).strip(),
        encoding="utf-8",
    )
    out = tmp_path / "candidate-profile.yaml"

    def broken_extract_pdf_text(path):
        raise ValueError("broken pdf")

    monkeypatch.setattr("job_research.cli.extract_pdf_text", broken_extract_pdf_text)

    result = CliRunner().invoke(app, ["build-profile", "--cv", str(cv), "--profile", str(profile), "--out", str(out)])

    assert result.exit_code == 1
    assert not out.exists()
    assert "CV input not readable" in result.stderr
    assert "broken pdf" in result.stderr
    assert "Traceback" not in result.stderr


def test_build_profile_reports_malformed_profile_markdown_cleanly(tmp_path) -> None:
    cv = tmp_path / "cv.txt"
    cv.write_text(
        dedent(
            """
            Tonio Example
            Location: France
            Languages: French, English
            Skills: Python
            """
        ).strip(),
        encoding="utf-8",
    )
    profile = tmp_path / "profile.md"
    profile.write_text(
        dedent(
            """
            # Candidate Profile

            ## Summary
            Junior data engineer.

            ## Target Roles
            Data Engineer

            ## Strengths
            - Python

            ## Skills To Emphasize
            - BigQuery

            ## Constraints
            - CDI only

            ## Notes
            - Slight preference for French listings.
            """
        ).strip(),
        encoding="utf-8",
    )
    out = tmp_path / "candidate-profile.yaml"

    result = CliRunner().invoke(app, ["build-profile", "--cv", str(cv), "--profile", str(profile), "--out", str(out)])

    assert result.exit_code == 1
    assert not out.exists()
    assert "profile markdown invalid" in result.stderr
    assert "Traceback" not in result.stderr