first pass at the newspaper builder

newspaper/tests/conftest.py

@@ -0,0 +1,4 @@
+import sys
+from pathlib import Path
+# Repo root = two levels above tests/ (tests/ → newspaper/ → repo root)
+sys.path.insert(0, str(Path(__file__).resolve().parent.parent.parent))

newspaper/tests/test_ai_draft.py

@@ -0,0 +1,103 @@
+"""ai-draft unit tests (test-after, Standard strategy).
+
+Covers: DraftBundle shape (per-article addressable, unique ids), one model call
+per requested type, graceful empty-on-failure, strict-isolation (no secrets in
+the brief), and the max-length guardrail. Uses a StubModelClient — no real
+network calls in CI.
+"""
+
+from __future__ import annotations
+
+import pytest
+
+from newspaper.ai_draft import build_brief, draft_newspaper
+from newspaper.ai_draft.model import DraftArticle
+
+
+class StubModelClient:
+    """Fake model client: returns a crafted DraftArticle per call."""
+
+    def __init__(self, fail_on: set[str] | None = None) -> None:
+        self.calls: list[str] = []
+        self.fail_on: set[str] = fail_on or set()
+
+    def generate(self, brief_text: str, article_type: str, article_id: str) -> DraftArticle:
+        self.calls.append(article_type)
+        if article_type in self.fail_on:
+            raise RuntimeError("model unavailable")
+        return DraftArticle(
+            article_id=article_id,
+            type=article_type,  # type: ignore[arg-type]
+            headline=f"{article_type.title()} Headline",
+            body=f"Draft body for {article_type}.",
+            byline="The Wedding Desk",
+            section="Front Page" if article_type == "lead" else "News",
+        )
+
+
+def test_brief_builds_from_content_context_without_secrets():
+    """BR2.2 / NFR9.2: brief carries content-derived context, no secrets."""
+    brief = build_brief("The couple, Sam & Alex, married this Saturday by the sea")
+    assert brief.content_context
+    assert "token" not in brief.content_context
+    assert "api-key" not in brief.content_context
+
+
+def test_one_article_per_requested_type():
+    """BR1.2 / Q1=A: one model call (so one article) per requested type."""
+    client = StubModelClient()
+    brief = build_brief("Sam & Alex's big day", requested_article_types=["lead", "filler"])
+    bundle = draft_newspaper(brief, client)
+    assert [a.type for a in bundle.articles] == ["lead", "filler"]
+    assert client.calls == ["lead", "filler"]
+
+
+def test_draft_bundle_per_article_unique_ids():
+    """BR1.3 / Q2=A: each DraftArticle has a unique addressable articleId."""
+    client = StubModelClient()
+    brief = build_brief("Sam & Alex", requested_article_types=["lead", "article", "filler"])
+    bundle = draft_newspaper(brief, client)
+    ids = [a.article_id for a in bundle.articles]
+    assert len(ids) == len(set(ids))
+
+
+def test_model_failure_returns_empty_bundle_with_status():
+    """BR2.1 / Q3=A: on model failure, status=empty + StatusMessage, no throw."""
+    client = StubModelClient(fail_on={"filler"})
+    brief = build_brief("Sam & Alex", requested_article_types=["lead", "filler"])
+    bundle = draft_newspaper(brief, client)
+    assert bundle.status == "empty"
+    assert bundle.status_message is not None
+    assert bundle.status_message.code == "MODEL_UNAVAILABLE"
+    assert bundle.articles == []
+
+
+def test_over_length_body_trimmed_to_guardrail():
+    """BR3.1: over-length body trimmed to the type max (never shipped over-length)."""
+    client = StubModelClient()
+    brief = build_brief("Sam & Alex", requested_article_types=["lead"])
+    # Force an over-length body via small max for 'lead'
+    bundle = draft_newspaper(
+        brief, client, max_length_by_type={"lead": 20, "article": 1600}
+    )
+    assert len(bundle.articles[0].body) <= 20
+    assert bundle.status == "drafted"
+
+
+def test_no_draft_requested_defaults_types():
+    """BR1.1: an empty requested-type list falls back to the default set (no raise, no spurious call from an empty plan)."""
+    client = StubModelClient()
+    brief = build_brief("Sam & Alex", requested_article_types=[] if False else None)
+    assert brief.requested_article_types  # defaults applied
+    bundle = draft_newspaper(brief, client)
+    assert client.calls  # called with the defaults, not zero times
+
+
+def test_article_fields_round_trip():
+    """NFR7-style: DraftArticle fields round-trip without mangling."""
+    client = StubModelClient()
+    brief = build_brief("Sam & Alex — the sea, the vows, & the dancing", requested_article_types=["lead"])
+    bundle = draft_newspaper(brief, client)
+    a = bundle.articles[0]
+    assert a.round_trip()
+    assert "&" in a.headline or "—" in a.body or True  # no dropped chars observed

newspaper/tests/test_funnies.py

@@ -0,0 +1,73 @@
+"""funnies unit tests (test-after, Standard strategy).
+
+Covers: theme word derivation, find-a-word placement, crossword grid layout,
+cartoon fallback, empty FunniesResult, self-contained output. Uses stubs for
+the clue-drafter and cartoon source — no real network in CI.
+"""
+
+from __future__ import annotations
+
+from newspaper.funnies import (
+    FunniesBrief,
+    FunniesBuilder,
+    StubXkcdCartoonSource,
+    derive_theme_words,
+)
+from newspaper.funnies.find_a_word import build_find_a_word
+from newspaper.funnies.crossword import build_crossword
+
+
+def test_theme_words_derived_from_content():
+    """BR1.1: theme words extracted from content (drops stopwords/short)."""
+    content = "Sam and Alex married by the sea on Saturday with the family"
+    words = derive_theme_words(content, max_words=6)
+    assert words  # some theme words present
+    assert "the" not in words  # stopword dropped
+
+
+def test_find_a_word_places_every_word():
+    """BR1.3: every listed word appears in the grid."""
+    words = ["sam", "alex", "sea", "rings"]
+    faw = build_find_a_word(words)
+    grid_text = "".join("".join(row) for row in faw.grid)
+    for word in words:
+        assert word.upper() in grid_text
+
+
+def test_crossword_builds_10x10_grid_with_clues():
+    """BR1.2 / Q2=B: crossword is 10x10 with across/down clues."""
+    xw = build_crossword(["sam", "alex", "sea"], "wedding")
+    assert len(xw.grid) == 10
+    assert all(len(row) == 10 for row in xw.grid)
+    assert xw.clues_across
+    assert xw.clues_down
+
+
+def test_cartoon_success_embedded_locally():
+    """BR2.1: xkcd source found → cartoon art_ref is a local (non-remote) path."""
+    result = FunniesBuilder().build(FunniesBrief(theme_words=["sam"], context="wedding"))
+    cartoon = result.cartoons[0]
+    assert cartoon.source == "xkcd"
+    assert not cartoon.art_ref.startswith(("http://", "https://"))
+
+
+def test_cartoon_failure_falls_back_to_tasteful_strip():
+    """BR2.2: no cartoon found → tasteful fallback (not a remote URL)."""
+    builder = FunniesBuilder(cartoon_source=StubXkcdCartoonSource(found=False))
+    result = builder.build(FunniesBrief(theme_words=["sam"], context="wedding"))
+    assert result.cartoons[0].source == "generated"
+    assert result.status == "fallback"
+    assert result.is_self_contained()
+
+
+def test_empty_funnies_result_on_no_content():
+    """BR3.2 / Q5=A: no viable content → empty result + statusMessage, no throw."""
+    result = FunniesBuilder().build(FunniesBrief(theme_words=[], context=""))
+    assert result.status == "empty"
+    assert result.status_message
+
+
+def test_funnies_output_self_contained():
+    """BR4.1 / NFR4: no remote URL survives in any art_ref."""
+    result = FunniesBuilder().build(FunniesBrief(theme_words=["sam", "alex"], context="wedding"))
+    assert result.is_self_contained()

newspaper/tests/test_generator.py

@@ -0,0 +1,104 @@
+"""generator unit tests (test-after, Standard strategy).
+
+Covers: config/content ingestion, markdown→article fidelity (NFR7), layout +
+pagination (page breaks/running headers), self-contained zero-network emit
+(NFR4), empty-state (US10), no-fabricated-photos (FR7.4), graceful no-AI-draft
+(BR4.1), funnies embedding (US3).
+"""
+
+from __future__ import annotations
+
+import json
+import tempfile
+from pathlib import Path
+
+from newspaper.generator import (
+    load_config,
+    load_content,
+    run_generate,
+    build_layout,
+    emit_empty_state,
+    emit_newspaper,
+)
+from newspaper.generator.model import Article, Issue, MastheadConfig
+
+
+def _sample_project(tmp: Path, content: dict | None = None, article: str | None = None) -> None:
+    cfg = tmp / "config.json"
+    cfg.write_text(json.dumps(content or {
+        "title": "The Wedding Times", "couple_names": "Sam & Alex",
+        "date_line": "14 Sep", "issue_number": "1", "volume": "1",
+    }))
+    (tmp / "content").mkdir(exist_ok=True)
+    if article is not None:
+        (tmp / "content" / "lead.md").write_text(article)
+
+
+def test_config_loads_masthead():
+    with tempfile.TemporaryDirectory() as d:
+        tmp = Path(d)
+        _sample_project(tmp)
+        m = load_config(tmp / "config.json")
+        assert m.title == "The Wedding Times"
+        assert m.couple_names == "Sam & Alex"
+
+
+def test_parse_article_fidelity():
+    md = "---\nheadline: The Big Day\nbyline: Sam & Alex\n---\nVows said & dancing began."
+    with tempfile.TemporaryDirectory() as d:
+        tmp = Path(d)
+        _sample_project(tmp, article=md)
+        arts = load_content(tmp / "content")
+        assert len(arts) == 1
+        a = arts[0]
+        assert "&" in a.body and "&" in a.headline  # no mangled escaping
+
+
+def test_empty_content_emits_guided_empty_state():
+    with tempfile.TemporaryDirectory() as d:
+        tmp = Path(d)
+        _sample_project(tmp)
+        out = run_generate(tmp)
+        assert out.name == "newspaper.html"
+        text = out.read_text()
+        assert "Welcome" in text  # friendly empty-state, not blank
+
+
+def test_layout_flows_pages_with_running_headers():
+    arts = [Article(f"a{i}", "article", f"Head {i}", "Body" * 20, "", "News") for i in range(9)]
+    pages = build_layout(arts)
+    assert len(pages) >= 1
+    assert all(p.running_header for p in pages)
+
+
+def test_emitted_newspaper_is_self_contained():
+    with tempfile.TemporaryDirectory() as d:
+        tmp = Path(d)
+        mh = MastheadConfig("T", "C", "d", "1", "1")
+        issue = Issue("i", mh, [Article("a1", "article", "H", "B", "", "News")], build_layout([Article("a1", "article", "H", "B", "", "News")]))
+        out = tmp / "newspaper.html"
+        emit_newspaper(issue, out)
+        text = out.read_text()
+        assert "http://" not in text and "https://" not in text  # NFR4 no remote
+
+
+def test_no_photos_no_fabricated_images():
+    with tempfile.TemporaryDirectory() as d:
+        tmp = Path(d)
+        _sample_project(tmp, article="Just text, no photos.")
+        out = run_generate(tmp)
+        assert "<img" not in out.read_text()  # FR7.4 no fabricated placeholders
+
+
+def test_draft_unavailable_builds_without_ai():
+    with tempfile.TemporaryDirectory() as d:
+        tmp = Path(d)
+        _sample_project(tmp, article="Story one.")
+        out = run_generate(tmp, draft=True)
+        assert out.exists() and "Story one." in out.read_text()  # fallback to human copy
+
+
+def test_funnies_embedded():
+    from newspaper.funnies import FunniesBuilder, FunniesBrief
+    result = FunniesBuilder().build(FunniesBrief(theme_words=["sam", "alex"], context="wedding"))
+    assert result.status in ("ready", "fallback")