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 | class SemanticSplitterNodeParser(NodeParser):
"""
Semantic node parser.
Splits a document into Nodes, with each node being a group of semantically related sentences.
Args:
buffer_size (int): number of sentences to group together when evaluating semantic similarity
embed_model: (BaseEmbedding): embedding model to use
sentence_splitter (Optional[Callable]): splits text into sentences
include_metadata (bool): whether to include metadata in nodes
include_prev_next_rel (bool): whether to include prev/next relationships
"""
sentence_splitter: SentenceSplitterCallable = Field(
default_factory=split_by_sentence_tokenizer,
description="The text splitter to use when splitting documents.",
exclude=True,
)
embed_model: SerializeAsAny[BaseEmbedding] = Field(
description="The embedding model to use to for semantic comparison",
)
buffer_size: int = Field(
default=1,
description=(
"The number of sentences to group together when evaluating semantic similarity. "
"Set to 1 to consider each sentence individually. "
"Set to >1 to group sentences together."
),
)
breakpoint_percentile_threshold: int = Field(
default=95,
description=(
"The percentile of cosine dissimilarity that must be exceeded between a "
"group of sentences and the next to form a node. The smaller this "
"number is, the more nodes will be generated"
),
)
@classmethod
def class_name(cls) -> str:
return "SemanticSplitterNodeParser"
@classmethod
def from_defaults(
cls,
embed_model: Optional[BaseEmbedding] = None,
breakpoint_percentile_threshold: Optional[int] = 95,
buffer_size: Optional[int] = 1,
sentence_splitter: Optional[Callable[[str], List[str]]] = None,
original_text_metadata_key: str = DEFAULT_OG_TEXT_METADATA_KEY,
include_metadata: bool = True,
include_prev_next_rel: bool = True,
callback_manager: Optional[CallbackManager] = None,
id_func: Optional[Callable[[int, Document], str]] = None,
) -> "SemanticSplitterNodeParser":
callback_manager = callback_manager or CallbackManager([])
sentence_splitter = sentence_splitter or split_by_sentence_tokenizer()
if embed_model is None:
try:
from llama_index.embeddings.openai import (
OpenAIEmbedding,
) # pants: no-infer-dep
embed_model = embed_model or OpenAIEmbedding()
except ImportError:
raise ImportError(
"`llama-index-embeddings-openai` package not found, "
"please run `pip install llama-index-embeddings-openai`"
)
id_func = id_func or default_id_func
return cls(
embed_model=embed_model,
breakpoint_percentile_threshold=breakpoint_percentile_threshold,
buffer_size=buffer_size,
sentence_splitter=sentence_splitter,
original_text_metadata_key=original_text_metadata_key,
include_metadata=include_metadata,
include_prev_next_rel=include_prev_next_rel,
callback_manager=callback_manager,
id_func=id_func,
)
def _parse_nodes(
self,
nodes: Sequence[BaseNode],
show_progress: bool = False,
**kwargs: Any,
) -> List[BaseNode]:
"""Parse document into nodes."""
all_nodes: List[BaseNode] = []
nodes_with_progress = get_tqdm_iterable(nodes, show_progress, "Parsing nodes")
for node in nodes_with_progress:
nodes = self.build_semantic_nodes_from_documents([node], show_progress)
all_nodes.extend(nodes)
return all_nodes
async def _aparse_nodes(
self,
nodes: Sequence[BaseNode],
show_progress: bool = False,
**kwargs: Any,
) -> List[BaseNode]:
"""Asynchronously parse document into nodes."""
all_nodes: List[BaseNode] = []
nodes_with_progress = get_tqdm_iterable(nodes, show_progress, "Parsing nodes")
for node in nodes_with_progress:
nodes = await self.abuild_semantic_nodes_from_documents(
[node], show_progress
)
all_nodes.extend(nodes)
return all_nodes
def build_semantic_nodes_from_documents(
self,
documents: Sequence[Document],
show_progress: bool = False,
) -> List[BaseNode]:
"""Build window nodes from documents."""
all_nodes: List[BaseNode] = []
for doc in documents:
text = doc.text
text_splits = self.sentence_splitter(text)
sentences = self._build_sentence_groups(text_splits)
combined_sentence_embeddings = self.embed_model.get_text_embedding_batch(
[s["combined_sentence"] for s in sentences],
show_progress=show_progress,
)
for i, embedding in enumerate(combined_sentence_embeddings):
sentences[i]["combined_sentence_embedding"] = embedding
distances = self._calculate_distances_between_sentence_groups(sentences)
chunks = self._build_node_chunks(sentences, distances)
nodes = build_nodes_from_splits(
chunks,
doc,
id_func=self.id_func,
)
all_nodes.extend(nodes)
return all_nodes
async def abuild_semantic_nodes_from_documents(
self,
documents: Sequence[Document],
show_progress: bool = False,
) -> List[BaseNode]:
"""Asynchronously build window nodes from documents."""
all_nodes: List[BaseNode] = []
for doc in documents:
text = doc.text
text_splits = self.sentence_splitter(text)
sentences = self._build_sentence_groups(text_splits)
combined_sentence_embeddings = (
await self.embed_model.aget_text_embedding_batch(
[s["combined_sentence"] for s in sentences],
show_progress=show_progress,
)
)
for i, embedding in enumerate(combined_sentence_embeddings):
sentences[i]["combined_sentence_embedding"] = embedding
distances = self._calculate_distances_between_sentence_groups(sentences)
chunks = self._build_node_chunks(sentences, distances)
nodes = build_nodes_from_splits(
chunks,
doc,
id_func=self.id_func,
)
all_nodes.extend(nodes)
return all_nodes
def _build_sentence_groups(
self, text_splits: List[str]
) -> List[SentenceCombination]:
sentences: List[SentenceCombination] = [
{
"sentence": x,
"index": i,
"combined_sentence": "",
"combined_sentence_embedding": [],
}
for i, x in enumerate(text_splits)
]
# Group sentences and calculate embeddings for sentence groups
for i in range(len(sentences)):
combined_sentence = ""
for j in range(i - self.buffer_size, i):
if j >= 0:
combined_sentence += sentences[j]["sentence"]
combined_sentence += sentences[i]["sentence"]
for j in range(i + 1, i + 1 + self.buffer_size):
if j < len(sentences):
combined_sentence += sentences[j]["sentence"]
sentences[i]["combined_sentence"] = combined_sentence
return sentences
def _calculate_distances_between_sentence_groups(
self, sentences: List[SentenceCombination]
) -> List[float]:
distances = []
for i in range(len(sentences) - 1):
embedding_current = sentences[i]["combined_sentence_embedding"]
embedding_next = sentences[i + 1]["combined_sentence_embedding"]
similarity = self.embed_model.similarity(embedding_current, embedding_next)
distance = 1 - similarity
distances.append(distance)
return distances
def _build_node_chunks(
self, sentences: List[SentenceCombination], distances: List[float]
) -> List[str]:
chunks = []
if len(distances) > 0:
breakpoint_distance_threshold = np.percentile(
distances, self.breakpoint_percentile_threshold
)
indices_above_threshold = [
i for i, x in enumerate(distances) if x > breakpoint_distance_threshold
]
# Chunk sentences into semantic groups based on percentile breakpoints
start_index = 0
for index in indices_above_threshold:
group = sentences[start_index : index + 1]
combined_text = "".join([d["sentence"] for d in group])
chunks.append(combined_text)
start_index = index + 1
if start_index < len(sentences):
combined_text = "".join(
[d["sentence"] for d in sentences[start_index:]]
)
chunks.append(combined_text)
else:
# If, for some reason we didn't get any distances (i.e. very, very small documents) just
# treat the whole document as a single node
chunks = [" ".join([s["sentence"] for s in sentences])]
return chunks
|
