alita-sdk 0.3.286__tar.gz → 0.3.288__tar.gz

{alita_sdk-0.3.286 → alita_sdk-0.3.288}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: alita_sdk
-Version: 0.3.286
+Version: 0.3.288
 Summary: SDK for building langchain agents using resources from Alita
 Author-email: Artem Rozumenko <artyom.rozumenko@gmail.com>, Mikalai Biazruchka <mikalai_biazruchka@epam.com>, Roman Mitusov <roman_mitusov@epam.com>, Ivan Krakhmaliuk <lifedjik@gmail.com>, Artem Dubrovskiy <ad13box@gmail.com>
 License-Expression: Apache-2.0
@@ -58,7 +58,7 @@ Requires-Dist: scipy==1.13.1; extra == "runtime"
 Requires-Dist: pytesseract==0.3.13; extra == "runtime"
 Requires-Dist: reportlab==4.2.5; extra == "runtime"
 Requires-Dist: svglib==1.5.1; extra == "runtime"
-Requires-Dist: rlpycairo==0.3.0; extra == "runtime"
+Requires-Dist: rlpycairo==0.4.0; extra == "runtime"
 Requires-Dist: cairocffi==1.7.1; extra == "runtime"
 Requires-Dist: docx2txt==0.8; extra == "runtime"
 Requires-Dist: mammoth==1.9.0; extra == "runtime"

{alita_sdk-0.3.286 → alita_sdk-0.3.288}/alita_sdk/runtime/langchain/document_loaders/constants.py

@@ -30,65 +30,76 @@ loaders_map = {
     '.png': {
         'class': AlitaImageLoader,
         'is_multimodal_processing': True,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     },
     '.jpg': {
         'class': AlitaImageLoader,
         'is_multimodal_processing': True,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     },
     '.jpeg': {
         'class': AlitaImageLoader,
         'is_multimodal_processing': True,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     },
     '.gif': {
         'class': AlitaImageLoader,
         'is_multimodal_processing': True,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     },
     '.bmp': {
         'class': AlitaImageLoader,
         'is_multimodal_processing': True,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     },
     '.svg': {
         'class': AlitaImageLoader,
         'is_multimodal_processing': True,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     },
     '.txt': {
         'class': AlitaTextLoader,
         'is_multimodal_processing': False,
         'kwargs': {
             'autodetect_encoding': True
-        }
+        },
+        'allowed_to_override': ['max_tokens']
     },
     '.yml': {
         'class': AlitaTextLoader,
         'is_multimodal_processing': False,
         'kwargs': {
             'autodetect_encoding': True
-        }
+        },
+        'allowed_to_override': ['max_tokens']
     },
     '.yaml': {
         'class': AlitaTextLoader,
         'is_multimodal_processing': False,
         'kwargs': {
             'autodetect_encoding': True
-        }
+        },
+        'allowed_to_override': ['max_tokens']
     },
     '.groovy': {
         'class': AlitaTextLoader,
         'is_multimodal_processing': False,
         'kwargs': {
             'autodetect_encoding': True
-        }
+        },
+        'allowed_to_override': ['max_tokens']
     },
     '.md': {
         'class': UnstructuredMarkdownLoader,
         'is_multimodal_processing': False,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     },
     '.csv': {
         'class': AlitaCSVLoader,
@@ -97,7 +108,8 @@ loaders_map = {
             'encoding': 'utf-8',
             'raw_content': False,
             'cleanse': False
-        }
+        },
+        'allowed_to_override': ['max_tokens']
     },
     '.xlsx': {
         'class': AlitaExcelLoader,
@@ -105,7 +117,8 @@ loaders_map = {
         'kwargs': {
             'raw_content': False,
             'cleanse': False
-        }
+        },
+        'allowed_to_override': ['max_tokens']
     },
     '.xls': {
         'class': AlitaExcelLoader,
@@ -113,57 +126,67 @@ loaders_map = {
         'kwargs': {
             'raw_content': False,
             'cleanse': False
-        }
+        },
+        'allowed_to_override': ['max_tokens']
     },
     '.pdf': {
         'class': AlitaPDFLoader,
         'is_multimodal_processing': False,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     },
     '.docx': {
         'class': AlitaDocxMammothLoader,
         'is_multimodal_processing': True,
         'kwargs': {
             'extract_images': True
-        }
+        },
+        'allowed_to_override': ['max_tokens', 'mode']
     },
     '.json': {
         'class': AlitaJSONLoader,
         'is_multimodal_processing': False,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     },
     '.jsonl': {
         'class': AirbyteJSONLoader,
         'is_multimodal_processing': False,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     },
     '.htm': {
         'class': UnstructuredHTMLLoader,
         'is_multimodal_processing': False,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     },
     '.html': {
         'class': UnstructuredHTMLLoader,
         'is_multimodal_processing': False,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     },
     '.ppt': {
         'class': AlitaPowerPointLoader,
         'is_multimodal_processing': False,
         'kwargs': {
             'mode': 'paged'
-        }
+        },
+        'allowed_to_override': ['max_tokens', 'mode']
     },
     '.pptx': {
         'class': AlitaPowerPointLoader,
         'is_multimodal_processing': False,
         'kwargs': {
             'mode': 'paged'
-        }
+        },
+        'allowed_to_override': ['max_tokens', 'mode']
     },
     '.py': {
         'class': PythonLoader,
         'is_multimodal_processing': False,
-        'kwargs': {}
+        'kwargs': {},
+        'allowed_to_override': ['max_tokens']
     }
 }

{alita_sdk-0.3.286 → alita_sdk-0.3.288}/alita_sdk/tools/utils/content_parser.py

@@ -203,7 +203,7 @@ def load_content_from_bytes(file_content: bytes, extension: str = None, loader_e
         if temp_file_path and os.path.exists(temp_file_path):
             os.remove(temp_file_path)
 
-def process_content_by_type(document: Document, content, extension_source: str, llm = None, chunking_config={}) -> Generator[Document, None, None]:
+def process_content_by_type(document: Document, content, extension_source: str, llm = None, chunking_config=None) -> Generator[Document, None, None]:
     temp_file_path = None
     try:
         extension = "." + extension_source.split('.')[-1].lower()
@@ -224,6 +224,16 @@ def process_content_by_type(document: Document, content, extension_source: str,
 
             loader_cls = loader_config['class']
             loader_kwargs = loader_config['kwargs']
+            # Determine which loader configuration keys are allowed to be overridden by user input.
+            # If 'allowed_to_override' is specified in the loader configuration, use it; otherwise, allow all keys in loader_kwargs.
+            allowed_to_override = loader_config.get('allowed_to_override', list(loader_kwargs.keys()))
+            # If a chunking_config is provided and contains custom configuration for the current file extension,
+            # update loader_kwargs with user-supplied values, but only for keys explicitly permitted in allowed_to_override.
+            # This ensures that only safe and intended parameters can be customized, preventing accidental or unauthorized changes
+            # to critical loader settings.
+            if chunking_config and (users_config_for_extension := chunking_config.get(extension, {})):
+                for key in set(users_config_for_extension.keys()) & set(allowed_to_override):
+                    loader_kwargs[key] = users_config_for_extension[key]
 
             loader = loader_cls(file_path=temp_file_path, **loader_kwargs)
             for chunk in loader.load():

{alita_sdk-0.3.286 → alita_sdk-0.3.288}/alita_sdk/tools/zephyr_enterprise/api_wrapper.py

@@ -157,12 +157,16 @@ class ZephyrApiWrapper(NonCodeIndexerToolkit):
         }
 
     def _base_loader(self, zql: str, **kwargs) -> Generator[Document, None, None]:
-        test_cases = self.get_testcases_by_zql(zql)
+        test_cases = self.get_testcases_by_zql(zql=zql, return_as_list=True)
         for test_case in test_cases:
             metadata = {
-                ("updated_on" if k == "lastModifiedOn" else "id" if k == "testcaseId" else k): str(v)
-                for k, v in test_case
-                if k != "id"
+                "updated_on": str(test_case.get("lastModifiedOn")),
+                "id": str(test_case.get("id")),
+                "name": test_case.get("name"),
+                "testcaseId": str(test_case.get("testcaseId")),
+                "projectId": test_case.get("projectId"),
+                "projectName": test_case.get("projectName"),
+                "testcaseType": test_case.get("testcaseType"),
             }
             yield Document(page_content='', metadata=metadata)
 
@@ -171,7 +175,7 @@ class ZephyrApiWrapper(NonCodeIndexerToolkit):
             try:
                 id = document.metadata['id']
                 test_case_content = self.get_test_case_steps(id)
-                document.page_content = test_case_content
+                document.page_content = f"Test case: {document.metadata['name']}\nTest_case_content: {test_case_content}"
             except Exception as e:
                 logging.error(f"Failed to process document: {e}")
             yield document

{alita_sdk-0.3.286 → alita_sdk-0.3.288}/alita_sdk.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: alita_sdk
-Version: 0.3.286
+Version: 0.3.288
 Summary: SDK for building langchain agents using resources from Alita
 Author-email: Artem Rozumenko <artyom.rozumenko@gmail.com>, Mikalai Biazruchka <mikalai_biazruchka@epam.com>, Roman Mitusov <roman_mitusov@epam.com>, Ivan Krakhmaliuk <lifedjik@gmail.com>, Artem Dubrovskiy <ad13box@gmail.com>
 License-Expression: Apache-2.0
@@ -58,7 +58,7 @@ Requires-Dist: scipy==1.13.1; extra == "runtime"
 Requires-Dist: pytesseract==0.3.13; extra == "runtime"
 Requires-Dist: reportlab==4.2.5; extra == "runtime"
 Requires-Dist: svglib==1.5.1; extra == "runtime"
-Requires-Dist: rlpycairo==0.3.0; extra == "runtime"
+Requires-Dist: rlpycairo==0.4.0; extra == "runtime"
 Requires-Dist: cairocffi==1.7.1; extra == "runtime"
 Requires-Dist: docx2txt==0.8; extra == "runtime"
 Requires-Dist: mammoth==1.9.0; extra == "runtime"

{alita_sdk-0.3.286 → alita_sdk-0.3.288}/alita_sdk.egg-info/requires.txt

@@ -63,7 +63,7 @@ scipy==1.13.1
 pytesseract==0.3.13
 reportlab==4.2.5
 svglib==1.5.1
-rlpycairo==0.3.0
+rlpycairo==0.4.0
 cairocffi==1.7.1
 docx2txt==0.8
 mammoth==1.9.0

{alita_sdk-0.3.286 → alita_sdk-0.3.288}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "alita_sdk"
-version = "0.3.286"
+version = "0.3.288"
 description = "SDK for building langchain agents using resources from Alita"
 readme = "README.md"
 requires-python = ">=3.10"
@@ -32,7 +32,7 @@ name = "Artem Dubrovskiy"
 email = "ad13box@gmail.com"
 
 [project.optional-dependencies]
-runtime = [ "langchain_core==0.3.49", "langchain<=0.3.22", "langchain_community~=0.3.7", "langchain-openai~=0.3.0", "langgraph-checkpoint-sqlite~=2.0.0", "langgraph-checkpoint-postgres==2.0.21", "langsmith>=0.3.45", "langgraph>=0.4.8,<0.5", "langchain_chroma~=0.2.2", "langchain-unstructured~=0.1.6", "langchain-postgres~=0.0.13", "keybert==0.8.3", "charset_normalizer==3.3.2", "unstructured[local-inference]==0.16.23", "unstructured_pytesseract==0.3.13", "unstructured_inference==0.8.7", "python-pptx==1.0.2", "pdf2image==1.16.3", "pikepdf==8.7.1", "pypdf==4.3.1", "pdfminer.six==20240706", "opencv-python==4.11.0.86", "python-docx==1.1.2", "openpyxl==3.1.2", "markdown==3.5.1", "beautifulsoup4==4.12.2", "gensim==4.3.3", "chromadb==0.5.20", "pgvector==0.2.5", "scipy==1.13.1", "pytesseract==0.3.13", "reportlab==4.2.5", "svglib==1.5.1", "rlpycairo==0.3.0", "cairocffi==1.7.1", "docx2txt==0.8", "mammoth==1.9.0", "opentelemetry-exporter-otlp-proto-grpc==1.25.0", "opentelemetry_api==1.25.0", "opentelemetry_instrumentation==0.46b0", "grpcio_status==1.63.0rc1", "protobuf==4.25.7", "sentence-transformers==2.7.0",]
+runtime = [ "langchain_core==0.3.49", "langchain<=0.3.22", "langchain_community~=0.3.7", "langchain-openai~=0.3.0", "langgraph-checkpoint-sqlite~=2.0.0", "langgraph-checkpoint-postgres==2.0.21", "langsmith>=0.3.45", "langgraph>=0.4.8,<0.5", "langchain_chroma~=0.2.2", "langchain-unstructured~=0.1.6", "langchain-postgres~=0.0.13", "keybert==0.8.3", "charset_normalizer==3.3.2", "unstructured[local-inference]==0.16.23", "unstructured_pytesseract==0.3.13", "unstructured_inference==0.8.7", "python-pptx==1.0.2", "pdf2image==1.16.3", "pikepdf==8.7.1", "pypdf==4.3.1", "pdfminer.six==20240706", "opencv-python==4.11.0.86", "python-docx==1.1.2", "openpyxl==3.1.2", "markdown==3.5.1", "beautifulsoup4==4.12.2", "gensim==4.3.3", "chromadb==0.5.20", "pgvector==0.2.5", "scipy==1.13.1", "pytesseract==0.3.13", "reportlab==4.2.5", "svglib==1.5.1", "rlpycairo==0.4.0", "cairocffi==1.7.1", "docx2txt==0.8", "mammoth==1.9.0", "opentelemetry-exporter-otlp-proto-grpc==1.25.0", "opentelemetry_api==1.25.0", "opentelemetry_instrumentation==0.46b0", "grpcio_status==1.63.0rc1", "protobuf==4.25.7", "sentence-transformers==2.7.0",]
 tools = [ "dulwich==0.21.6", "paramiko==3.3.1", "pygithub==2.3.0", "python-gitlab==4.5.0", "gitpython==3.1.43", "atlassian-python-api~=3.41", "atlassian_python_api==3.41.16", "jira==3.8.0", "qtest-swagger-client==0.0.3", "testrail-api==1.13.2", "azure-devops==7.1.0b4", "msrest==0.7.1", "python-graphql-client~=0.4.3", "zephyr-python-api==0.1.0", "pyral==1.6.0", "boto3>=1.37.23", "azure-core==1.30.2", "azure-identity==1.16.0", "azure-keyvault-keys==4.9.0", "azure-keyvault-secrets==4.8.0", "azure-mgmt-core==1.4.0", "azure-mgmt-resource==23.0.1", "azure-mgmt-storage==21.1.0", "azure-storage-blob==12.23.1", "azure-search-documents==11.5.2", "PyMySQL==1.1.1", "psycopg2-binary==2.9.10", "Office365-REST-Python-Client==2.5.14", "python-docx==1.1.2", "python-pptx==1.0.2", "pypdf2~=3.0.1", "FigmaPy==2018.1.0", "pandas==2.2.3", "factor_analyzer==0.5.1", "statsmodels==0.14.4", "tabulate==0.9.0", "tree_sitter==0.20.2", "tree-sitter-languages==1.10.2", "astor~=0.8.1", "markdownify~=1.1.0", "requests_openapi==1.0.5", "duckduckgo_search==5.3.0", "playwright>=1.52.0", "google-api-python-client==2.154.0", "wikipedia==1.4.0", "lxml==5.2.2", "beautifulsoup4", "pymupdf==1.24.9", "googlemaps==4.10.0", "yagmail==0.15.293", "pysnc==1.1.10", "shortuuid==1.0.13", "yarl==1.17.1", "langmem==0.0.27", "textract-py3==2.1.1", "slack_sdk==3.35.0", "deltalake==1.0.2", "google_cloud_bigquery==3.34.0",]
 community = [ "retry-extended==0.2.3", "pyobjtojson==0.3", "elitea-analyse==0.1.2",]
 all = [ "alita-sdk[runtime]", "alita-sdk[tools]", "alita-sdk[community]",]