Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
|
@@ -167,28 +167,6 @@ else:
|
|
| 167 |
embedding_function=sentence_transformer_ef
|
| 168 |
)
|
| 169 |
|
| 170 |
-
dataframe: pl.DataFrame = pl.scan_parquet(
|
| 171 |
-
"hf://datasets/louisbrulenaudet/lemone-docs-embeded/data/train-00000-of-00001.parquet"
|
| 172 |
-
).filter(
|
| 173 |
-
pl.col(
|
| 174 |
-
"text"
|
| 175 |
-
).is_not_null()
|
| 176 |
-
).collect()
|
| 177 |
-
|
| 178 |
-
collection.add(
|
| 179 |
-
embeddings=dataframe["lemone_pro_embeddings"].to_list(),
|
| 180 |
-
documents=dataframe["text"].to_list(),
|
| 181 |
-
metadatas=dataframe.drop(
|
| 182 |
-
[
|
| 183 |
-
"lemone_pro_embeddings",
|
| 184 |
-
"text"
|
| 185 |
-
]
|
| 186 |
-
).to_dicts(),
|
| 187 |
-
ids=[
|
| 188 |
-
str(i) for i in range(0, dataframe.shape[0])
|
| 189 |
-
]
|
| 190 |
-
)
|
| 191 |
-
|
| 192 |
|
| 193 |
def trim_input_ids(
|
| 194 |
input_ids,
|
|
|
|
| 167 |
embedding_function=sentence_transformer_ef
|
| 168 |
)
|
| 169 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 170 |
|
| 171 |
def trim_input_ids(
|
| 172 |
input_ids,
|