File size: 3,081 Bytes
660dde6 dbabef2 660dde6 dbabef2 660dde6 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 | from __future__ import annotations
from tqdm import tqdm
from qdrant_client.models import (
PointStruct
)
from db.parsers.bsa.embedder import (
LegalEmbedder
)
from db.parsers.bsa.qdrant_store import (
QdrantStore
)
class LegalIngestionPipeline:
def __init__(
self,
collection_name: str
):
self.embedder = (
LegalEmbedder()
)
self.store = (
QdrantStore(
collection_name=
collection_name
)
)
# =====================================================
# INGEST
# =====================================================
def ingest(
self,
chunks,
batch_size: int = 64,
recreate_collection: bool = False
):
# ==========================================
# VECTOR SIZE
# ==========================================
vector_size = (
self.embedder.dimension
)
# ==========================================
# RECREATE COLLECTION
# ==========================================
if recreate_collection:
self.store.delete_collection()
# ==========================================
# CREATE COLLECTION
# ==========================================
self.store.create_collection(
vector_size
)
# ==========================================
# UPSERT BATCHES
# ==========================================
point_id = 1
for start in tqdm(
range(
0,
len(chunks),
batch_size
),
desc="Indexing"
):
batch = chunks[
start:
start + batch_size
]
texts = [
chunk[
"enriched_text"
]
for chunk in batch
]
embeddings = (
self.embedder.embed(
texts
)
)
points = []
for (
chunk,
embedding
) in zip(
batch,
embeddings
):
points.append(
PointStruct(
id=point_id,
vector=
embedding.tolist(),
payload=
chunk
)
)
point_id += 1
self.store.upsert(
points
)
# ==========================================
# SUMMARY
# ==========================================
print()
print(
f"Indexed "
f"{len(chunks)} chunks "
f"into "
f"{self.store.collection_name}"
)
print(
f"Total points: "
f"{self.store.count()}"
) |