start project

tests/test_log_parser.py

@@ -0,0 +1,50 @@
+"""Log parser tests (Chapter 12: highest testing priority per Ch12)."""
+from __future__ import annotations
+
+from app.services.log_parser.apache import combined_parser, common_parser
+from app.services.log_parser.format_compiler import compile_format, parse_apache_timestamp
+
+COMMON_LINE = '127.0.0.1 - - [10/Oct/2026:13:55:36 -0700] "GET /index.html HTTP/1.1" 200 1234\n'
+COMBINED_LINE = (
+    '127.0.0.1 - - [10/Oct/2026:13:55:36 -0700] "GET /index.html HTTP/1.1" 200 1234 '
+    '"https://example.com/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) Chrome/120.0"\n'
+)
+MALFORMED_LINE = "this is not a log line\n"
+
+
+def test_common_parser_parses_valid_line():
+    entry = common_parser().parse_line(COMMON_LINE)
+    assert entry is not None
+    assert entry.ip == "127.0.0.1"
+    assert entry.method == "GET"
+    assert entry.path == "/index.html"
+    assert entry.status_code == 200
+    assert entry.bytes_sent == 1234
+    assert entry.referrer is None
+    assert entry.user_agent is None
+
+
+def test_combined_parser_parses_referrer_and_user_agent():
+    entry = combined_parser().parse_line(COMBINED_LINE)
+    assert entry is not None
+    assert entry.referrer == "https://example.com/"
+    assert "Chrome" in entry.user_agent
+
+
+def test_malformed_line_returns_none_not_raises():
+    assert common_parser().parse_line(MALFORMED_LINE) is None
+
+
+def test_custom_format_string_compiles_and_parses():
+    custom_format = '%h %t "%r" %>s'
+    compiled = compile_format(custom_format)
+    match = compiled.pattern.match('127.0.0.1 [10/Oct/2026:13:55:36 -0700] "GET /x HTTP/1.1" 404')
+    assert match is not None
+    assert match.group("ip") == "127.0.0.1"
+    assert match.group("status") == "404"
+
+
+def test_parse_apache_timestamp_normalizes_to_utc():
+    dt = parse_apache_timestamp("10/Oct/2026:13:55:36 -0700")
+    assert dt.hour == 20  # -0700 -> UTC
+    assert dt.tzinfo is None