refactor app_gui.py for mainloop. Added tests and CI

Author: SculptTechProject

.github/workflows/build_binary.yml

@@ -0,0 +1,50 @@
+---
+name: CI
+on: [push, pull_request]
+
+permissions:
+  contents: write
+
+jobs:
+  lint-test:
+    runs-on: ubuntu-24.04
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: '3.11'
+
+      - name: Cache pip
+        uses: actions/cache@v4
+        with:
+          path: ~/.cache/pip
+          key: ${{ runner.os }}-pip-${{ hashFiles('requirements.txt') }}
+
+      - name: Install system GUI deps
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y python3-tk xvfb
+
+      - name: Install deps
+        run: |
+          python -m pip install --upgrade pip
+          pip install -r requirements.txt
+          pip install isort
+
+      - name: Auto-format (black + isort)
+        run: |
+          isort .
+          black .
+
+      - name: Auto-commit formatted files
+        uses: stefanzweifel/git-auto-commit-action@v5
+        with:
+          commit_message: "ci(fmt): black+isort [skip ci]"
+          file_pattern: "**/*.py"
+
+      - name: Run tests (headless Tk)
+        env:
+          SERPAPI_KEY: "test-key"
+        run: xvfb-run -a pytest -q -m "not slow"

.github/workflows/ci.yml

@@ -135,31 +135,23 @@ def choose_dir():
 
 
 # --- UI ---
-root = tk.Tk()
-root.title("GinioCrawler")
-root.geometry("560x220")
-
-# fraza
-tk.Label(root, text="Fraza do wyszukania:").pack(anchor="w", padx=12, pady=(12, 0))
-entry_query = tk.Entry(root)
-entry_query.pack(fill="x", padx=12, pady=6)
-entry_query.focus()
-
-# folder wyjściowy
-frm = tk.Frame(root)
-frm.pack(fill="x", padx=12, pady=(0, 6))
-tk.Label(frm, text="Folder wyjściowy:").pack(side="left")
-out_dir_var = tk.StringVar(value=str((Path.cwd() / "wyniki")))
-entry_dir = tk.Entry(frm, textvariable=out_dir_var)
-entry_dir.pack(side="left", fill="x", expand=True, padx=(8, 6))
-tk.Button(frm, text="Wybierz…", command=choose_dir).pack(side="left")
-
-# start + status
-btn_start = tk.Button(root, text="Start", command=start)
-btn_start.pack(padx=12, pady=6)
-prog = ttk.Progressbar(root, mode="indeterminate")
-prog.pack(fill="x", padx=12, pady=(4, 8))
-status = tk.StringVar(value="Gotowy")
-tk.Label(root, textvariable=status, anchor="w").pack(fill="x", padx=12, pady=(0, 8))
-
-root.mainloop()
+def build_ui():
+    global root, entry_query, btn_start, prog, status, out_dir_var
+    root = tk.Tk()
+    root.title("GinioCrawler")
+    root.geometry("560x220")
+    tk.Label(root, text="Fraza do wyszukania:").pack(anchor="w", padx=12, pady=(12, 0))
+    entry_query = tk.Entry(root); entry_query.pack(fill="x", padx=12, pady=6); entry_query.focus()
+    frm = tk.Frame(root); frm.pack(fill="x", padx=12, pady=(0, 6))
+    tk.Label(frm, text="Folder wyjściowy:").pack(side="left")
+    out_dir_var = tk.StringVar(value=str((Path.cwd() / "wyniki")))
+    entry_dir = tk.Entry(frm, textvariable=out_dir_var); entry_dir.pack(side="left", fill="x", expand=True, padx=(8, 6))
+    tk.Button(frm, text="Wybierz…", command=choose_dir).pack(side="left")
+    btn_start = tk.Button(root, text="Start", command=start); btn_start.pack(padx=12, pady=6)
+    prog = ttk.Progressbar(root, mode="indeterminate"); prog.pack(fill="x", padx=12, pady=(4, 8))
+    status = tk.StringVar(value="Gotowy"); tk.Label(root, textvariable=status, anchor="w").pack(fill="x", padx=12, pady=(0, 8))
+    return root
+
+if __name__ == "__main__":
+    build_ui()
+    root.mainloop()

.github/workflows/license.yml

@@ -150,7 +150,7 @@ async def run(query):
     try:
         import os, sys, csv, asyncio
         from pathlib import Path
-        from datetime import datetime  # ważne: klasa datetime, nie moduł
+        from datetime import datetime
 
         q = input("Podaj szukane słowo: ").strip()
         if not q:

app_gui.py

@@ -24,4 +24,6 @@ freezegun
 
 dirty-equals
 
-pytest-cov   
+pytest-cov
+
+pytest

main.py

@@ -0,0 +1,4 @@
+[pytest]
+addopts = -q
+markers =
+  slow: wolniejsze testy sieci/Playwright

requirements.txt

@@ -0,0 +1,32 @@
+import httpx, pytest, respx
+from main import crawl_one, run
+
+@pytest.mark.asyncio
+@respx.mock
+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"])
+
+@pytest.mark.asyncio
+@respx.mock
+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"}

tests/.gitkeep

@@ -0,0 +1,17 @@
+from pathlib import Path
+import pandas as pd
+from main import save_results
+
+def test_save_results_creates_files(tmp_path, monkeypatch):
+    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
+    xlsx_called = {}
+    def fake_write(csv, xlsx):
+        xlsx_called["csv"] = csv; xlsx_called["xlsx"] = xlsx
+        return real_write_excel(csv, xlsx)
+    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"}