ben/vath
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

remove hyphen for underscore in nomenclature, remove dependency

Browse Source
This commit is contained in:
eulaly
2026-05-05 16:47:11 -04:00
parent fd9d656e13
commit 683bfb324f
5 changed files with 67 additions and 127 deletions
Download Patch File Download Diff File Expand all files Collapse all files

87
tests/test_gpt4o_analysis.py → tests/analysis_gpt4o_realtime.py
View File

@@ -1,15 +1,14 @@
"""Unit tests for analysis/gpt4o/analysis.py — no real API calls."""
"""Unit tests for analysis/gpt4o/analysis_realtime.py — no real API calls."""
import json
import sys
from pathlib import Path
from unittest.mock import MagicMock, patch
from unittest.mock import MagicMock
import pytest
# Make the module importable without installing as a package
sys.path.insert(0, str(Path(__file__).parent.parent / "analysis" / "gpt4o"))
import analysis as gpt4o
import analysis_realtime as rt
# ---------------------------------------------------------------------------
@@ -51,26 +50,25 @@ def _mock_client(response_content: str = MOCK_RESPONSE_CONTENT):
# Prompt versioning
def test_prompt_version_is_7_hex_chars():
assert len(gpt4o.PROMPT_VERSION) == 7
assert all(c in "0123456789abcdef" for c in gpt4o.PROMPT_VERSION)
assert len(rt.PROMPT_VERSION) == 7
assert all(c in "0123456789abcdef" for c in rt.PROMPT_VERSION)
def test_prompt_version_changes_with_system_prompt():
def test_prompt_version_matches_prompt_file():
import hashlib
alt = hashlib.sha256(("CHANGED" + gpt4o.USER_TEMPLATE).encode("utf-8")).hexdigest()[:7]
assert alt != gpt4o.PROMPT_VERSION
prompt_file = Path(__file__).parent.parent / "analysis" / "prompt-1.txt"
expected = hashlib.sha256(prompt_file.read_text(encoding="utf-8").strip().encode()).hexdigest()[:7]
assert rt.PROMPT_VERSION == expected
def test_prompt_version_is_stable():
import hashlib
v2 = hashlib.sha256(
(gpt4o.SYSTEM_PROMPT + gpt4o.USER_TEMPLATE).encode("utf-8")
).hexdigest()[:7]
assert v2 == gpt4o.PROMPT_VERSION
v2 = hashlib.sha256(rt.SYSTEM_PROMPT.encode("utf-8")).hexdigest()[:7]
assert v2 == rt.PROMPT_VERSION
# ---------------------------------------------------------------------------
# Item detection via load_items
# load_items
def test_load_items_separates_forum_and_comments(tmp_path):
jsonl = tmp_path / "test.jsonl"
@@ -78,7 +76,7 @@ def test_load_items_separates_forum_and_comments(tmp_path):
json.dumps(FORUM_ITEM) + "\n" + json.dumps(COMMENT_ITEM) + "\n",
encoding="utf-8",
)
forum, comments = gpt4o.load_items(jsonl)
forum, comments = rt.load_items(jsonl)
assert forum is not None
assert forum["reg_title"] == FORUM_ITEM["reg_title"]
assert len(comments) == 1
@@ -88,18 +86,15 @@ def test_load_items_separates_forum_and_comments(tmp_path):
def test_load_items_no_forum(tmp_path):
jsonl = tmp_path / "test.jsonl"
jsonl.write_text(json.dumps(COMMENT_ITEM) + "\n", encoding="utf-8")
forum, comments = gpt4o.load_items(jsonl)
forum, comments = rt.load_items(jsonl)
assert forum is None
assert len(comments) == 1
def test_load_items_skips_blank_lines(tmp_path):
jsonl = tmp_path / "test.jsonl"
jsonl.write_text(
"\n" + json.dumps(COMMENT_ITEM) + "\n\n",
encoding="utf-8",
)
_, comments = gpt4o.load_items(jsonl)
jsonl.write_text("\n" + json.dumps(COMMENT_ITEM) + "\n\n", encoding="utf-8")
_, comments = rt.load_items(jsonl)
assert len(comments) == 1
@@ -108,40 +103,37 @@ def test_load_items_skips_blank_lines(tmp_path):
def test_truncation_applied():
long_comment = {**COMMENT_ITEM, "text": "x" * 7000}
messages, truncated = gpt4o.build_messages(long_comment, FORUM_ITEM)
messages, truncated = rt.build_messages(long_comment, FORUM_ITEM)
assert truncated is True
user_content = messages[1]["content"]
assert "... [truncated]" in user_content
# The x's in the prompt must not exceed MAX_COMMENT_CHARS
x_count = user_content.count("x")
assert x_count == gpt4o.MAX_COMMENT_CHARS
assert "... [truncated]" in messages[1]["content"]
assert messages[1]["content"].count("x") == rt.MAX_COMMENT_CHARS
def test_no_truncation_for_short_comment():
_, truncated = gpt4o.build_messages(COMMENT_ITEM, FORUM_ITEM)
_, truncated = rt.build_messages(COMMENT_ITEM, FORUM_ITEM)
assert truncated is False
def test_empty_text_fallback():
empty = {**COMMENT_ITEM, "text": ""}
messages, truncated = gpt4o.build_messages(empty, FORUM_ITEM)
messages, truncated = rt.build_messages(empty, FORUM_ITEM)
assert "[No body text provided]" in messages[1]["content"]
assert truncated is False
def test_none_text_fallback():
none_text = {**COMMENT_ITEM, "text": None}
messages, _ = gpt4o.build_messages(none_text, FORUM_ITEM)
messages, _ = rt.build_messages(none_text, FORUM_ITEM)
assert "[No body text provided]" in messages[1]["content"]
def test_missing_forum_uses_unknown_context():
messages, _ = gpt4o.build_messages(COMMENT_ITEM, None)
messages, _ = rt.build_messages(COMMENT_ITEM, None)
assert "[unknown]" in messages[1]["content"]
def test_reg_context_included_in_prompt():
messages, _ = gpt4o.build_messages(COMMENT_ITEM, FORUM_ITEM)
messages, _ = rt.build_messages(COMMENT_ITEM, FORUM_ITEM)
assert FORUM_ITEM["reg_title"] in messages[1]["content"]
assert "HB 145" in messages[1]["content"]
@@ -150,8 +142,7 @@ def test_reg_context_included_in_prompt():
# Output record schema
def test_output_record_all_keys_present():
client = _mock_client()
record = gpt4o.analyze_comment(client, COMMENT_ITEM, FORUM_ITEM, "run-123", "gpt-4o")
record = rt.analyze_comment(_mock_client(), COMMENT_ITEM, FORUM_ITEM, "run-123", "gpt-4o")
required = {
"run_id", "forum_id", "comment_id", "analyzed_at", "model", "prompt_version",
"stance", "stance_confidence", "stance_rationale", "tone", "tags",
@@ -161,8 +152,7 @@ def test_output_record_all_keys_present():
def test_output_record_correct_types():
client = _mock_client()
record = gpt4o.analyze_comment(client, COMMENT_ITEM, FORUM_ITEM, "run-123", "gpt-4o")
record = rt.analyze_comment(_mock_client(), COMMENT_ITEM, FORUM_ITEM, "run-123", "gpt-4o")
assert record["stance"] == "support"
assert isinstance(record["stance_confidence"], float)
assert isinstance(record["tags"], list)
@@ -171,13 +161,12 @@ def test_output_record_correct_types():
def test_output_record_metadata():
client = _mock_client()
record = gpt4o.analyze_comment(client, COMMENT_ITEM, FORUM_ITEM, "run-123", "gpt-4o")
record = rt.analyze_comment(_mock_client(), COMMENT_ITEM, FORUM_ITEM, "run-123", "gpt-4o")
assert record["run_id"] == "run-123"
assert record["forum_id"] == "452"
assert record["comment_id"] == "87914"
assert record["model"] == "gpt-4o"
assert record["prompt_version"] == gpt4o.PROMPT_VERSION
assert record["prompt_version"] == rt.PROMPT_VERSION
assert record["input_title"] == COMMENT_ITEM["title"]
@@ -185,12 +174,12 @@ def test_output_record_metadata():
# Error handling
def test_error_record_on_api_failure():
client = MagicMock()
import openai as _openai
client = MagicMock()
client.chat.completions.create.side_effect = _openai.RateLimitError(
"rate limit", response=MagicMock(status_code=429), body={}
)
record = gpt4o.analyze_comment(client, COMMENT_ITEM, FORUM_ITEM, "run-123", "gpt-4o")
record = rt.analyze_comment(client, COMMENT_ITEM, FORUM_ITEM, "run-123", "gpt-4o")
assert record["error"] is not None
assert record["stance"] is None
assert record["tone"] is None
@@ -198,8 +187,7 @@ def test_error_record_on_api_failure():
def test_error_record_on_bad_json():
client = _mock_client("not valid json{{{")
record = gpt4o.analyze_comment(client, COMMENT_ITEM, FORUM_ITEM, "run-123", "gpt-4o")
record = rt.analyze_comment(_mock_client("not valid json{{{"), COMMENT_ITEM, FORUM_ITEM, "run-123", "gpt-4o")
assert record["error"] is not None
assert record["stance"] is None
@@ -210,21 +198,18 @@ def test_error_record_on_bad_json():
def test_run_id_is_shared_across_records():
client = _mock_client()
run_id = "fixed-run-id"
r1 = gpt4o.analyze_comment(client, COMMENT_ITEM, FORUM_ITEM, run_id, "gpt-4o")
r2 = gpt4o.analyze_comment(client, {**COMMENT_ITEM, "comment_id": "99999"}, FORUM_ITEM, run_id, "gpt-4o")
r1 = rt.analyze_comment(client, COMMENT_ITEM, FORUM_ITEM, run_id, "gpt-4o")
r2 = rt.analyze_comment(client, {**COMMENT_ITEM, "comment_id": "99999"}, FORUM_ITEM, run_id, "gpt-4o")
assert r1["run_id"] == r2["run_id"] == run_id
# ---------------------------------------------------------------------------
# Filename parsing
# Filename helpers
def test_scrape_ts_extracted_from_filename():
p = Path("output/forum452_comments_2026-05-05T17-33-54+00-00.jsonl")
ts = gpt4o._scrape_ts_from_filename(p)
assert ts == "2026-05-05T17-33-54+00-00"
assert rt._scrape_ts_from_filename(p) == "2026-05-05T17-33-54+00-00"
def test_scrape_ts_fallback_for_unknown_filename():
p = Path("output/somefile.jsonl")
ts = gpt4o._scrape_ts_from_filename(p)
assert ts == "unknown"
assert rt._scrape_ts_from_filename(Path("output/somefile.jsonl")) == "unknown"

0
tests/test_forum_spider.py → tests/scrape_forum_spider.py
View File