claude
fda572f59e
mcp 2.x already on origin/main (#10). Does not change BM25-first search_docs — n=6 is too small to flip the default. - Numbered [1] citations via docs_mcp/format.py - Eval P@1 + JSONL sidecar + eval.pvalue + eval.trace - Heading-recursive chunker, keep chunk-0 and MAX_CHARS=4000 - Nomic prefixes at embed time only; stored text unprefixed Closes #12
42 lines
1.3 KiB
Python
42 lines
1.3 KiB
Python
"""Prefix helpers — no Ollama, no Chroma."""
|
|
from __future__ import annotations
|
|
|
|
import unittest
|
|
|
|
from rag.embeddings import apply_prefix, embed_texts
|
|
|
|
|
|
class FakeEmbedder:
|
|
def __init__(self) -> None:
|
|
self.seen: list[str] = []
|
|
|
|
def __call__(self, texts: list[str]) -> list[list[float]]:
|
|
self.seen.extend(texts)
|
|
return [[0.0, 1.0] for _ in texts]
|
|
|
|
|
|
class PrefixTests(unittest.TestCase):
|
|
def test_index_sends_doc_prefix_keeps_raw_text(self) -> None:
|
|
fake = FakeEmbedder()
|
|
raw = ["hello"]
|
|
vecs = embed_texts(raw, prefix="search_document: ", ef=fake)
|
|
self.assertEqual(fake.seen, ["search_document: hello"])
|
|
self.assertEqual(raw, ["hello"]) # caller storage unchanged
|
|
self.assertEqual(len(vecs), 1)
|
|
|
|
def test_query_sends_query_prefix(self) -> None:
|
|
fake = FakeEmbedder()
|
|
embed_texts(["q"], prefix="search_query: ", ef=fake)
|
|
self.assertEqual(fake.seen, ["search_query: q"])
|
|
self.assertNotIn("q", fake.seen)
|
|
|
|
def test_empty_prefix_is_raw(self) -> None:
|
|
self.assertEqual(apply_prefix("hello", ""), "hello")
|
|
fake = FakeEmbedder()
|
|
embed_texts(["hello"], prefix="", ef=fake)
|
|
self.assertEqual(fake.seen, ["hello"])
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|