File size: 11,218 Bytes
2f044c1 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 |
import csv
import json
import pickle
from pathlib import Path
from typing import Dict, List, Union
from relik.common.log import get_logger
logger = get_logger(__name__)
class Document:
def __init__(
self,
text: str,
id: int | None = None,
metadata: Dict | None = None,
**kwargs,
):
self.text = text
# if id is not provided, we use the hash of the text
self.id = id if id is not None else hash(text)
# if metadata is not provided, we use an empty dictionary
self.metadata = metadata or {}
def __str__(self):
return f"{self.id}:{self.text}"
def __repr__(self):
return self.__str__()
def __eq__(self, other):
if isinstance(other, Document):
return self.id == other.id
elif isinstance(other, int):
return self.id == other
elif isinstance(other, str):
return self.text == other
else:
raise ValueError(
f"Document must be compared with a Document, an int or a str, got `{type(other)}`"
)
def to_dict(self):
return {"text": self.text, "id": self.id, "metadata": self.metadata}
@classmethod
def from_dict(cls, d: Dict):
return cls(**d)
@classmethod
def from_file(cls, file_path: Union[str, Path], **kwargs):
with open(file_path, "r") as f:
d = json.load(f)
return cls.from_dict(d)
def save(self, file_path: Union[str, Path], **kwargs):
with open(file_path, "w") as f:
json.dump(self.to_dict(), f, indent=2)
class DocumentStore:
"""
A document store is a collection of documents.
Args:
documents (:obj:`List[Document]`):
The documents to store.
"""
def __init__(self, documents: List[Document] = None) -> None:
if documents is None:
documents = []
# if self.ingore_case:
# documents = [doc.lower() for doc in documents]
self._documents = documents
# build an index for the documents
self._documents_index = {doc.id: doc for doc in self._documents}
# build a reverse index for the documents
self._documents_reverse_index = {doc.text: doc for doc in self._documents}
def __len__(self):
return len(self._documents)
def __getitem__(self, index):
return self._documents[index]
def __iter__(self):
return iter(self._documents)
def __contains__(self, item):
if isinstance(item, int):
return item in self._documents_index
elif isinstance(item, str):
return item in self._documents_reverse_index
elif isinstance(item, Document):
return item.id in self._documents_index
# return item in self._documents_index
def __str__(self):
return f"DocumentStore with {len(self)} documents"
def __repr__(self):
return self.__str__()
def get_document_from_id(self, id: int) -> Document | None:
"""
Retrieve a document by its ID.
Args:
id (`int`):
The ID of the document to retrieve.
Returns:
Optional[Document]: The document with the given ID, or None if it does not exist.
"""
if id not in self._documents_index:
logger.warning(f"Document with id `{id}` does not exist, skipping")
return self._documents_index.get(id, None)
def get_document_from_text(self, text: str) -> Document | None:
"""
Retrieve the document by its text.
Args:
text (`str`):
The text of the document to retrieve.
Returns:
Optional[Document]: The document with the given text, or None if it does not exist.
"""
if text not in self._documents_reverse_index:
logger.warning(f"Document with text `{text}` does not exist, skipping")
return self._documents_reverse_index.get(text, None)
def add_documents(self, documents: List[Document] | List[Dict]) -> List[Document]:
"""
Add a list of documents to the document store.
Args:
documents (`List[Document]`):
The documents to add.
Returns:
List[Document]: The documents just added.
"""
return [
self.add_document(doc)
if isinstance(doc, Document)
else self.add_document(Document.from_dict(doc))
for doc in documents
]
def add_document(
self,
text: str | Document,
id: int | None = None,
metadata: Dict | None = None,
) -> Document:
"""
Add a document to the document store.
Args:
text (`str`):
The text of the document to add.
id (`int`, optional, defaults to None):
The ID of the document to add.
metadata (`Dict`, optional, defaults to None):
The metadata of the document to add.
Returns:
Document: The document just added.
"""
if isinstance(text, str):
if id is None:
# get the len of the documents and add 1
id = len(self._documents) # + 1
text = Document(text, id, metadata)
if text in self:
logger.warning(f"Document {text} already exists, skipping")
return self._documents_index[text.id]
self._documents.append(text)
self._documents_index[text.id] = text
self._documents_reverse_index[text.text] = text
return text
# if id in self._documents_index:
# logger.warning(f"Document with id `{id}` already exists, skipping")
# return self._documents_index[id]
# if text_or_document in self._documents_reverse_index:
# logger.warning(f"Document with text `{text_or_document}` already exists, skipping")
# return self._documents_reverse_index[text_or_document]
# self._documents.append(Document(text_or_document, id, metadata))
# self._documents_index[id] = self._documents[-1]
# self._documents_reverse_index[text_or_document] = self._documents[-1]
# return self._documents_index[id]
def delete_document(self, document: int | str | Document) -> bool:
"""
Delete a document from the document store.
Args:
document (`int`, `str` or `Document`):
The document to delete.
Returns:
bool: True if the document has been deleted, False otherwise.
"""
if isinstance(document, int):
return self.delete_by_id(document)
elif isinstance(document, str):
return self.delete_by_text(document)
elif isinstance(document, Document):
return self.delete_by_document(document)
else:
raise ValueError(
f"Document must be an int, a str or a Document, got `{type(document)}`"
)
def delete_by_id(self, id: int) -> bool:
"""
Delete a document by its ID.
Args:
id (`int`):
The ID of the document to delete.
Returns:
bool: True if the document has been deleted, False otherwise.
"""
if id not in self._documents_index:
logger.warning(f"Document with id `{id}` does not exist, skipping")
return False
del self._documents_reverse_index[self._documents_index[id]]
del self._documents_index[id]
return True
def delete_by_text(self, text: str) -> bool:
"""
Delete a document by its text.
Args:
text (`str`):
The text of the document to delete.
Returns:
bool: True if the document has been deleted, False otherwise.
"""
if text not in self._documents_reverse_index:
logger.warning(f"Document with text `{text}` does not exist, skipping")
return False
del self._documents_reverse_index[text]
del self._documents_index[self._documents_index[text]]
return True
def delete_by_document(self, document: Document) -> bool:
"""
Delete a document by its text.
Args:
document (:obj:`Document`):
The document to delete.
Returns:
bool: True if the document has been deleted, False otherwise.
"""
if document.id not in self._documents_index:
logger.warning(f"Document {document} does not exist, skipping")
return False
del self._documents[self._documents.index(document)]
del self._documents_index[document.id]
del self._documents_reverse_index[self._documents_index[document.id]]
def to_dict(self):
return [doc.to_dict() for doc in self._documents]
@classmethod
def from_dict(cls, d):
return cls([Document.from_dict(doc) for doc in d])
@classmethod
def from_file(cls, file_path: Union[str, Path], **kwargs):
with open(file_path, "r") as f:
# load a json lines file
d = [Document.from_dict(json.loads(line)) for line in f]
return cls(d)
@classmethod
def from_pickle(cls, file_path: Union[str, Path], **kwargs):
with open(file_path, "rb") as handle:
d = pickle.load(handle)
return cls(d)
@classmethod
def from_tsv(
cls,
file_path: Union[str, Path],
ingore_case: bool = False,
delimiter: str = "\t",
**kwargs,
):
d = []
# load a tsv/csv file and take the header into account
# the header must be `id\ttext\t[list of metadata keys]`
with open(file_path, "r", encoding="utf8") as f:
csv_reader = csv.reader(f, delimiter=delimiter, **kwargs)
header = next(csv_reader)
id, text, *metadata_keys = header
for i, row in enumerate(csv_reader):
# check if id can be casted to int
# if not, we add it to the metadata and use `i` as id
try:
s_id = int(row[header.index(id)])
row_metadata_keys = metadata_keys
except ValueError:
row_metadata_keys = [id] + metadata_keys
s_id = i
d.append(
Document(
text=row[header.index(text)].strip().lower()
if ingore_case
else row[header.index(text)].strip(),
id=s_id, # row[header.index(id)],
metadata={
key: row[header.index(key)] for key in row_metadata_keys
},
)
)
return cls(d)
def save(self, file_path: Union[str, Path], **kwargs):
with open(file_path, "w") as f:
for doc in self._documents:
# save as json lines
f.write(json.dumps(doc.to_dict()) + "\n")
|