fixed tests

Author: SculptTechProject

requirements.txt

@@ -26,4 +26,16 @@ dirty-equals
 
 pytest-cov
 
-pytest
+pytest
+
+anyio
+
+pytest-asyncio
+
+pytest-tornasync
+
+pytest-trio
+
+pytest-twisted
+
+twisted

src/main.py

@@ -12,12 +12,15 @@
 
 
 def write_excel(csv_path, xlsx_path):
-    df = pd.read_csv(csv_path)
+    import pandas as pd
+    from openpyxl.styles import Font
+
+    df = pd.read_csv(csv_path, dtype=str, keep_default_na=False)
+
     for col in ("emails", "phones"):
         if col in df.columns:
             df[col] = (
                 df[col]
-                .fillna("")
                 .str.replace(";", " ")
                 .str.replace(r"\s+", " ", regex=True)
                 .str.strip()
@@ -26,21 +29,13 @@ def write_excel(csv_path, xlsx_path):
     with pd.ExcelWriter(xlsx_path, engine="openpyxl") as w:
         df.to_excel(w, index=False, sheet_name="Wyniki")
         ws = w.sheets["Wyniki"]
-
-        # nagłówki, freeze, filtr
         for c in ws[1]:
             c.font = Font(bold=True)
         ws.freeze_panes = "A2"
         ws.auto_filter.ref = ws.dimensions
-
-        # auto-szerokość (z limitem)
         for col in ws.columns:
             length = max(len(str(c.value)) if c.value else 0 for c in col)
-            ws.column_dimensions[col[0].column_letter].width = min(
-                max(12, int(length * 0.9)), 60
-            )
-
-        # hiperlinki
+            ws.column_dimensions[col[0].column_letter].width = min(max(12, int(length * 0.9)), 60)
         cols = [c for c in ("url", "contact_url") if c in df.columns]
         for row in range(2, ws.max_row + 1):
             for name in cols:
@@ -52,6 +47,7 @@ def write_excel(csv_path, xlsx_path):
                     cell.style = "Hyperlink"
 
 
+
 def in_robots(url: str) -> bool:
     base = f"{urllib.parse.urlsplit(url).scheme}://{urllib.parse.urlsplit(url).netloc}"
     robots = rp.RobotFileParser()

tests/conftest.py

@@ -1,4 +1,6 @@
-import os, sys
-ROOT = os.path.abspath(os.path.join(os.path.dirname(__file__), '..', 'src'))
-if ROOT not in sys.path:
-    sys.path.insert(0, ROOT)
+import sys
+from pathlib import Path
+
+REPO_ROOT = Path(__file__).resolve().parents[1]
+if str(REPO_ROOT) not in sys.path:
+    sys.path.insert(0, str(REPO_ROOT))

tests/test_crawl_run.py

@@ -1,61 +1,27 @@
-<<<<<<< HEAD
-import httpx, pytest, respx
+import httpx, respx, pytest
 from src.main import crawl_one, run
-=======
-import httpx
-import pytest
-import respx
 
-from main import crawl_one, run
->>>>>>> e772c4167010570baed9b34907b8a8834fe77e81
-
-
-@pytest.mark.asyncio
 @respx.mock
+@pytest.mark.anyio
 async def test_crawl_one_merges_contact(monkeypatch):
-    monkeypatch.setattr("ginio.in_robots", lambda url: True)
-    respx.get("https://site.test/").mock(
-        return_value=httpx.Response(
-            200,
-            text="""
-        <a href="/contact">Contact</a>
-        <p>[email protected]</p>
-    """,
-        )
-    )
-    respx.get("https://site.test/contact").mock(
-        return_value=httpx.Response(
-            200,
-            text="""
-        <p>[email protected] 123 456 789</p>
-    """,
-        )
-    )
-    async with httpx.AsyncClient() as client:
-        out = await crawl_one("https://site.test/", client)
-    assert "[email protected]" in out["emails"] and "[email protected]" in out["emails"]
-    assert any("123" in p for p in out["phones"])
-
+    monkeypatch.setattr("src.main.in_robots", lambda u: True)
+    respx.get("https://site.test/").mock(return_value=httpx.Response(200, text="""
+        <a href="/contact">Contact</a><p>[email protected]</p>"""))
+    respx.get("https://site.test/contact").mock(return_value=httpx.Response(200, text="""
+        <p>[email protected] 123 456 789</p>"""))
+    async with httpx.AsyncClient() as c:
+        out = await crawl_one("https://site.test/", c)
+    assert {"[email protected]","[email protected]"} <= set(out["emails"])
 
-@pytest.mark.asyncio
 @respx.mock
+@pytest.mark.anyio
 async def test_run_full(monkeypatch):
-    respx.get("https://serpapi.com/search").mock(
-        return_value=httpx.Response(
-            200,
-            json={
-                "organic_results": [{"link": "https://a.pl"}, {"link": "https://b.pl"}]
-            },
-        )
-    )
-    # stub robots + strony
-    monkeypatch.setattr("ginio.in_robots", lambda url: True)
-    respx.get("https://a.pl").mock(
-        return_value=httpx.Response(200, text="<title>A</title>")
-    )
-    respx.get("https://b.pl").mock(
-        return_value=httpx.Response(200, text="<title>B</title>")
-    )
-    monkeypatch.setenv("SERPAPI_KEY", "x")
-    results = await run("foo")
-    assert {r["title"] for r in results} == {"A", "B"}
+    respx.get("https://serpapi.com/search").mock(return_value=httpx.Response(
+        200, json={"organic_results":[{"link":"https://a.pl"},{"link":"https://b.pl"}]}
+    ))
+    monkeypatch.setattr("src.main.in_robots", lambda u: True)
+    respx.get("https://a.pl").mock(return_value=httpx.Response(200, text="<title>A</title>"))
+    respx.get("https://b.pl").mock(return_value=httpx.Response(200, text="<title>B</title>"))
+    monkeypatch.setenv("SERPAPI_KEY","x")
+    out = await run("foo")
+    assert {r["title"] for r in out} == {"A","B"}

tests/test_excel.py

@@ -1,42 +1,21 @@
 from pathlib import Path
-
 import pandas as pd
-<<<<<<< HEAD
-from src.main import save_results
-=======
-
-from main import save_results
->>>>>>> e772c4167010570baed9b34907b8a8834fe77e81
-
+from src.app_gui import save_results
+from src.main import write_excel as real_write_excel
 
 def test_save_results_creates_files(tmp_path, monkeypatch):
-<<<<<<< HEAD
-    root_out = tmp_path/"wyniki"
-    data = [{"url":"https://x","title":"X","emails":["a@x"],"phones":["123"],"contact_url":None}]
-    from src.main import write_excel as real_write_excel
-=======
     root_out = tmp_path / "wyniki"
-    data = [
-        {
-            "url": "https://x",
-            "title": "X",
-            "emails": ["a@x"],
-            "phones": ["123"],
-            "contact_url": None,
-        }
-    ]
-    from main import write_excel as real_write_excel
+    data = [{"url":"https://x","title":"X","emails":["a@x"],"phones":["123"],"contact_url":None}]
 
->>>>>>> e772c4167010570baed9b34907b8a8834fe77e81
-    xlsx_called = {}
+    called = {}
+    def fake_write(csv_path, xlsx_path):
+        called["csv"] = csv_path; called["xlsx"] = xlsx_path
+        return real_write_excel(csv_path, xlsx_path)
 
-    def fake_write(csv, xlsx):
-        xlsx_called["csv"] = csv
-        xlsx_called["xlsx"] = xlsx
-        return real_write_excel(csv, xlsx)
+    monkeypatch.setattr("src.app_gui.write_excel", fake_write)
 
-    monkeypatch.setattr("ginio.write_excel", fake_write)
     csv_p, xlsx_p = save_results(data, "20250101_000000", root_out)
     assert Path(csv_p).exists() and Path(xlsx_p).exists()
     df = pd.read_excel(xlsx_p)
-    assert set(df.columns) >= {"url", "title", "emails", "phones", "contact_url"}
+    assert {"url","title","emails","phones","contact_url"}.issubset(df.columns)
+    assert called["csv"] == str(csv_p) and called["xlsx"] == str(xlsx_p)

tests/test_gui.py

@@ -1,59 +1,36 @@
 from unittest.mock import patch
 
-
 def test_ensure_api_key_env(monkeypatch):
-<<<<<<< HEAD
     monkeypatch.setenv("SERPAPI_KEY","secret")
-    from src.main import ensure_api_key
-    assert ensure_api_key()=="secret"
-
-def test_ensure_api_key_prompt(tmp_path, monkeypatch):
-    monkeypatch.setenv("APPDATA", str(tmp_path))
-    from src.main import ensure_api_key
-=======
-    monkeypatch.setenv("SERPAPI_KEY", "secret")
-    from main import ensure_api_key
-
+    from src.app_gui import ensure_api_key
     assert ensure_api_key() == "secret"
 
-
 def test_ensure_api_key_prompt(tmp_path, monkeypatch):
+    monkeypatch.delenv("SERPAPI_KEY", raising=False)
     monkeypatch.setenv("APPDATA", str(tmp_path))
-    from main import ensure_api_key
-
->>>>>>> e772c4167010570baed9b34907b8a8834fe77e81
-    with patch("ginio.simpledialog.askstring", return_value="abc"):
-        key = ensure_api_key()
-    assert key == "abc"
-
+    from src.app_gui import ensure_api_key
+    with patch("src.app_gui.simpledialog.askstring", return_value="abc"):
+        assert ensure_api_key() == "abc"
 
 def test_start_calls_run_without_threading(monkeypatch):
-<<<<<<< HEAD
     import src.app_gui as g
-=======
-    import app_gui as g
-
->>>>>>> e772c4167010570baed9b34907b8a8834fe77e81
     g.build_ui()
-    g.entry_query.delete(0, "end")
-    g.entry_query.insert(0, "kawa")
-    monkeypatch.setenv("SERPAPI_KEY", "x")
+    g.entry_query.delete(0,'end'); g.entry_query.insert(0,"kawa")
+    monkeypatch.setenv("SERPAPI_KEY","x")
 
     class DummyThread:
-        def __init__(self, target, daemon):
-            self.target = target
-
-        def start(self):
-            self.target()
-
-    monkeypatch.setattr("ginio_gui.threading.Thread", DummyThread)
-    monkeypatch.setattr(
-        "ginio_gui.run",
-        lambda q: [
-            {"url": "u", "title": "t", "emails": [], "phones": [], "contact_url": None}
-        ],
-    )
+        def __init__(self, target, daemon): self.target = target
+        def start(self): self.target()
+
+    monkeypatch.setattr("src.app_gui.threading.Thread", DummyThread)
+    monkeypatch.setattr(g.root, "after", lambda ms, fn: fn(), raising=False)
+
+    async def fake_run(q):
+        return [{"url":"u","title":"t","emails":[],"phones":[],"contact_url":None}]
+    monkeypatch.setattr("src.app_gui.run", fake_run)
+
     with patch.object(g, "messagebox"):
         g.start()
         assert g.btn_start["state"] == "normal"
         assert "OK — zapisano" in g.status.get()
+

tests/test_net.py

@@ -1,30 +1,18 @@
-<<<<<<< HEAD
-import httpx, pytest, respx
+import httpx, respx, pytest
 from src.main import fetch
-=======
-import httpx
-import pytest
-import respx
 
-from main import fetch
->>>>>>> e772c4167010570baed9b34907b8a8834fe77e81
-
-
-@pytest.mark.asyncio
 @respx.mock
+@pytest.mark.anyio
 async def test_fetch_ok(monkeypatch):
-    monkeypatch.setattr("ginio.in_robots", lambda url: True)
-    respx.get("https://x.test/ok").mock(
-        return_value=httpx.Response(200, text="<h1>ok</h1>")
-    )
-    async with httpx.AsyncClient() as client:
-        html = await fetch("https://x.test/ok", client)
+    monkeypatch.setattr("src.main.in_robots", lambda u: True)
+    respx.get("https://x.test/ok").mock(return_value=httpx.Response(200, text="<h1>ok</h1>"))
+    async with httpx.AsyncClient() as c:
+        html = await fetch("https://x.test/ok", c)
     assert "ok" in html
 
-
-@pytest.mark.asyncio
 @respx.mock
+@pytest.mark.anyio
 async def test_fetch_respects_robots(monkeypatch):
-    monkeypatch.setattr("ginio.in_robots", lambda url: False)
-    async with httpx.AsyncClient() as client:
-        assert await fetch("https://x.test/nope", client) is None
+    monkeypatch.setattr("src.main.in_robots", lambda u: False)
+    async with httpx.AsyncClient() as c:
+        assert await fetch("https://x.test/nope", c) is None

tests/test_parse.py

@@ -1,14 +1,8 @@
-<<<<<<< HEAD
 from src.main import parse_info, absolutize
-=======
-from main import absolutize, parse_info
-
->>>>>>> e772c4167010570baed9b34907b8a8834fe77e81
 
 def test_absolutize():
     assert absolutize("https://ex.com/dir/", "../a") == "https://ex.com/a"
 
-
 def test_parse_info_extracts_contact():
     html = """
     <html><head><title> ACME </title></head>