cognee 0.3.8__py3-none-any.whl → 0.3.9__py3-none-any.whl

cognee/infrastructure/files/utils/get_file_metadata.py

@@ -1,6 +1,6 @@
 import io
 import os.path
-from typing import BinaryIO, TypedDict
+from typing import BinaryIO, TypedDict, Optional
 from pathlib import Path
 
 from cognee.shared.logging_utils import get_logger
@@ -27,7 +27,7 @@ class FileMetadata(TypedDict):
     file_size: int
 
 
-async def get_file_metadata(file: BinaryIO) -> FileMetadata:
+async def get_file_metadata(file: BinaryIO, name: Optional[str] = None) -> FileMetadata:
     """
     Retrieve metadata from a file object.
 
@@ -53,7 +53,7 @@ async def get_file_metadata(file: BinaryIO) -> FileMetadata:
     except io.UnsupportedOperation as error:
         logger.error(f"Error retrieving content hash for file: {file.name} \n{str(error)}\n\n")
 
-    file_type = guess_file_type(file)
+    file_type = guess_file_type(file, name)
 
     file_path = getattr(file, "name", None) or getattr(file, "full_name", None)
 

cognee/infrastructure/files/utils/guess_file_type.py

@@ -1,6 +1,9 @@
-from typing import BinaryIO
+import io
+from pathlib import Path
+from typing import BinaryIO, Optional, Any
 import filetype
-from .is_text_content import is_text_content
+from tempfile import SpooledTemporaryFile
+from filetype.types.base import Type
 
 
 class FileTypeException(Exception):
@@ -22,7 +25,7 @@ class FileTypeException(Exception):
         self.message = message
 
 
-def guess_file_type(file: BinaryIO) -> filetype.Type:
+def guess_file_type(file: BinaryIO, name: Optional[str] = None) -> filetype.Type:
     """
     Guess the file type from the given binary file stream.
 
@@ -39,12 +42,23 @@ def guess_file_type(file: BinaryIO) -> filetype.Type:
 
         - filetype.Type: The guessed file type, represented as filetype.Type.
     """
+
+    # Note: If file has .txt or .text extension, consider it a plain text file as filetype.guess may not detect it properly
+    # as it contains no magic number encoding
+    ext = None
+    if isinstance(file, str):
+        ext = Path(file).suffix
+    elif name is not None:
+        ext = Path(name).suffix
+
+    if ext in [".txt", ".text"]:
+        file_type = Type("text/plain", "txt")
+        return file_type
+
     file_type = filetype.guess(file)
 
     # If file type could not be determined consider it a plain text file as they don't have magic number encoding
     if file_type is None:
-        from filetype.types.base import Type
-
         file_type = Type("text/plain", "txt")
 
     if file_type is None:

cognee/infrastructure/loaders/core/audio_loader.py

@@ -42,6 +42,7 @@ class AudioLoader(LoaderInterface):
             "audio/wav",
             "audio/amr",
             "audio/aiff",
+            "audio/x-wav",
         ]
 
     @property

cognee/modules/ingestion/data_types/BinaryData.py

@@ -30,7 +30,7 @@ class BinaryData(IngestionData):
 
     async def ensure_metadata(self):
         if self.metadata is None:
-            self.metadata = await get_file_metadata(self.data)
+            self.metadata = await get_file_metadata(self.data, name=self.name)
 
             if self.metadata["name"] is None:
                 self.metadata["name"] = self.name

{cognee-0.3.8.dist-info → cognee-0.3.9.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cognee
-Version: 0.3.8
+Version: 0.3.9
 Summary: Cognee - is a library for enriching LLM context with a semantic layer for better understanding and reasoning.
 Project-URL: Homepage, https://www.cognee.ai
 Project-URL: Repository, https://github.com/topoteretes/cognee

{cognee-0.3.8.dist-info → cognee-0.3.9.dist-info}/RECORD

@@ -246,8 +246,8 @@ cognee/infrastructure/files/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeR
 cognee/infrastructure/files/utils/extract_text_from_file.py,sha256=-v0uvK6nXP6Q2Ia0GjIi97WntPFX6sWZQXO_Fg9TrCc,1112
 cognee/infrastructure/files/utils/get_data_file_path.py,sha256=Xz9anl6yYxK6wETKhVeK4f3ahjw58Aj8YkyJkJONOvc,1549
 cognee/infrastructure/files/utils/get_file_content_hash.py,sha256=0L_wgsRF8zqmtisFWcp4agDs7WovvBjiVWNQ_NCPKwo,1338
-cognee/infrastructure/files/utils/get_file_metadata.py,sha256=WpyOTUf2CPFT8ZlxOWuchg34xu8HVrsMP7cpahsFX7g,2292
-cognee/infrastructure/files/utils/guess_file_type.py,sha256=kPLubfG2XXr0FZTepWNSq5zn2W3-sPASK8F_1CCR9OE,1444
+cognee/infrastructure/files/utils/get_file_metadata.py,sha256=3U0usuzEuGbVY0PBqQl9FyU1fWeHzlc8DRqNWZaSoc8,2336
+cognee/infrastructure/files/utils/guess_file_type.py,sha256=s1CkS4DhhxyYIxlapGfVZKOQUbDwselljLRvlrP4BvE,1923
 cognee/infrastructure/files/utils/is_text_content.py,sha256=iNZWCECNLMjlQfOQAujVQis7prA1cqsscRRSQsxccZo,1316
 cognee/infrastructure/files/utils/open_data_file.py,sha256=3TPsTUDCH6SOuvbwNembE-YRiFDhb9yCqOC537b6iGY,2155
 cognee/infrastructure/llm/LLMGateway.py,sha256=o_XXoj6qbTb2zO9MCxut81CkZODJUMtRbLAS854JSzY,2478
@@ -367,7 +367,7 @@ cognee/infrastructure/loaders/get_loader_engine.py,sha256=cPJefAHFAWU1HXQoWqCpwh
 cognee/infrastructure/loaders/supported_loaders.py,sha256=LBCvqk6PGJsTtFB5vUpArPmxSegRr81y1oNAejzn1mE,961
 cognee/infrastructure/loaders/use_loader.py,sha256=ncfUFVohPox296m8tMeIl6Hnk1xRvHcpRCmwZXKPZ1s,598
 cognee/infrastructure/loaders/core/__init__.py,sha256=LTr8FWDXpG-Oxp8nwwn0KnHT97aIK6_FWiswmy7g40Q,230
-cognee/infrastructure/loaders/core/audio_loader.py,sha256=rNXsCuLCIinNOaZZXw778bp3ptUMSoLiBaq9cTZ4NFI,3015
+cognee/infrastructure/loaders/core/audio_loader.py,sha256=VBiJb3tgog51yP14CMLQP3RdF6cmj5dCsHeAHDMlCz8,3042
 cognee/infrastructure/loaders/core/image_loader.py,sha256=b8etveiidIvCw7PXqM2ldyxXDhkqi4-Ak-4BbX664Is,3390
 cognee/infrastructure/loaders/core/text_loader.py,sha256=zkFhjm_QeQu4fWv_Wkoe0O1Kpe9_uBgskkjeWn0sV-M,2991
 cognee/infrastructure/loaders/external/__init__.py,sha256=UwLJK81I1Atuw3FN34EDy8NKe7sltxRLZiONYHfoW4o,884
@@ -483,7 +483,7 @@ cognee/modules/ingestion/discover_directory_datasets.py,sha256=wtqYoZ5MpGc_FuzyK
 cognee/modules/ingestion/get_matched_datasets.py,sha256=BL2H_3t3wDWqcJxlo6uv-1u__g2E5OMwJYFsLCSDF34,475
 cognee/modules/ingestion/identify.py,sha256=4-oD_VjdJC9oUmJjuLJ1a6BX1-GKbw-rNgWyB9GyhC8,346
 cognee/modules/ingestion/save_data_to_file.py,sha256=SZFrWbkRCvENQ05JXAAKZgcVm4-s795ZPnhCgdGM5HY,1230
-cognee/modules/ingestion/data_types/BinaryData.py,sha256=E9B6N9nJQd8uG2IUJctrWa2y1QkC4txNBAAl9a9aP2g,1060
+cognee/modules/ingestion/data_types/BinaryData.py,sha256=UUo3MZdGaIePs0jbI_Nwwwr6HBnsdtqwt0FJU1d6rqw,1076
 cognee/modules/ingestion/data_types/IngestionData.py,sha256=JLKzItByitgfQAeEo7-qaRRce_weij-t3YY_nJ4wFy0,309
 cognee/modules/ingestion/data_types/S3BinaryData.py,sha256=Kdd4R2anhhIPQZ-5xihcWrMPY_MPHIfS4GJYP4ZeraU,1805
 cognee/modules/ingestion/data_types/TextData.py,sha256=HpIgFqFHm66D-_bgEljUSsh4GSfsLaOj_ubFc_RalNQ,939
@@ -942,9 +942,9 @@ distributed/tasks/queued_add_edges.py,sha256=kz1DHE05y-kNHORQJjYWHUi6Q1QWUp_v3Dl
 distributed/tasks/queued_add_nodes.py,sha256=aqK4Ij--ADwUWknxYpiwbYrpa6CcvFfqHWbUZW4Kh3A,452
 distributed/workers/data_point_saving_worker.py,sha256=kmaQy2A2J7W3k9Gd5lyoiT0XYOaJmEM8MbkKVOFOQVU,4729
 distributed/workers/graph_saving_worker.py,sha256=b5OPLLUq0OBALGekdp73JKxU0GrMlVbO4AfIhmACKkQ,4724
-cognee-0.3.8.dist-info/METADATA,sha256=GI-zL_URnvcCPoJI7XcApWndzMacGJ3ULKLzyxR27lE,14938
-cognee-0.3.8.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-cognee-0.3.8.dist-info/entry_points.txt,sha256=fAozOD9Vs4kgYwRhBiZoLCIXu-OSZqVxKGv45l19uok,88
-cognee-0.3.8.dist-info/licenses/LICENSE,sha256=pHHjSQj1DD8SDppW88MMs04TPk7eAanL1c5xj8NY7NQ,11344
-cognee-0.3.8.dist-info/licenses/NOTICE.md,sha256=6L3saP3kSpcingOxDh-SGjMS8GY79Rlh2dBNLaO0o5c,339
-cognee-0.3.8.dist-info/RECORD,,
+cognee-0.3.9.dist-info/METADATA,sha256=CdDPzdK2sCof1lzKjzfD2Fbeqyl2ObVE8EuGwxxKCiQ,14938
+cognee-0.3.9.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+cognee-0.3.9.dist-info/entry_points.txt,sha256=fAozOD9Vs4kgYwRhBiZoLCIXu-OSZqVxKGv45l19uok,88
+cognee-0.3.9.dist-info/licenses/LICENSE,sha256=pHHjSQj1DD8SDppW88MMs04TPk7eAanL1c5xj8NY7NQ,11344
+cognee-0.3.9.dist-info/licenses/NOTICE.md,sha256=6L3saP3kSpcingOxDh-SGjMS8GY79Rlh2dBNLaO0o5c,339
+cognee-0.3.9.dist-info/RECORD,,