ws-bom-robot-app 0.0.80__py3-none-any.whl → 0.0.81__py3-none-any.whl

ws_bom_robot_app/llm/vector_store/generator.py

@@ -1,137 +1,137 @@
-import os, gc, shutil, logging, traceback
-import asyncio, aiofiles, aiofiles.os
-from fastapi import HTTPException
-from fastapi.responses import StreamingResponse
-from langchain_core.documents import Document
-from ws_bom_robot_app.llm.vector_store.loader.base import Loader
-from ws_bom_robot_app.llm.models.api import RulesRequest, KbRequest, VectorDbResponse
-from ws_bom_robot_app.llm.vector_store.integration.manager import IntegrationManager
-from ws_bom_robot_app.llm.vector_store.db.manager import VectorDbManager
-from ws_bom_robot_app.config import config
-from ws_bom_robot_app.llm.models.kb import load_endpoints
-from ws_bom_robot_app.llm.utils.download import download_files
-
-async def _cleanup_directory(directory_path: str):
-  if os.path.exists(directory_path):
-    await asyncio.to_thread(shutil.rmtree, directory_path)
-
-#@timer
-async def rules(rq: RulesRequest) -> VectorDbResponse:
-  _config = rq.config()
-  db_name = rq.out_name()
-  store_path = os.path.join(_config.robot_data_folder, _config.robot_data_db_folder, _config.robot_data_db_folder_store, db_name)
-  try:
-    await VectorDbManager.get_strategy(rq.vector_type).create(rq.embeddings(),[Document(page_content=rule, metadata={"source": "rules"}) for rule in rq.rules], store_path) #type: ignore
-    db_file_path = shutil.make_archive(os.path.join(_config.robot_data_folder, _config.robot_data_db_folder, _config.robot_data_db_folder_out, db_name), "zip", store_path)
-    return VectorDbResponse(file = os.path.basename(db_file_path), vector_type=rq.vector_type)
-  except Exception as e:
-    try:
-      await _cleanup_directory(store_path)
-    finally:
-      return VectorDbResponse(success = False, error = str(e))
-  finally:
-    gc.collect()
-
-#@atimer
-async def kb(rq: KbRequest) -> VectorDbResponse:
-  os.environ['MPLCONFIGDIR'] = './tmp/.matplotlib'
-  _config = rq.config()
-  db_name = rq.out_name()
-  src_path = os.path.join(_config.robot_data_folder, _config.robot_data_db_folder, _config.robot_data_db_folder_src)
-  working_path = os.path.join(src_path, db_name)
-
-  if all([not rq.files,not rq.endpoints,not rq.integrations]):
-    return VectorDbResponse(success = False, error = "No files, endpoints or integrations provided")
-  else:
-    await aiofiles.os.makedirs(src_path, exist_ok=True)
-    await aiofiles.os.makedirs(working_path, exist_ok=True)
-
-  documents: list[Document] = []
-  # Download/copy all files
-  if rq.files:
-    try:
-      loaders = Loader(working_path)
-      filter_file_extensions = loaders.managed_file_extensions()
-      files_to_download = [file for file in rq.files if not os.path.exists(os.path.join(src_path, os.path.basename(file)))]
-      if files_to_download:
-        await download_files(
-          [f"{_config.robot_cms_host}/{_config.robot_cms_kb_folder}/{os.path.basename(file)}" for file in files_to_download if any([file.endswith(ext) for ext in filter_file_extensions])],
-          src_path, authorization=_config.robot_cms_auth)
-      # copy files to working tmp folder
-      for file in rq.files:
-        async with aiofiles.open(os.path.join(src_path, os.path.basename(file)), 'rb') as src_file:
-          async with aiofiles.open(os.path.join(working_path, os.path.basename(file)), 'wb') as dest_file:
-            await dest_file.write(await src_file.read())
-      #load files
-      try:
-        documents.extend(await loaders.load())
-      except Exception as e:
-        tb = traceback.format_exc()
-        _error = f"File loader failure: {e} | {tb}"
-        logging.warning(_error)
-        return VectorDbResponse(success = False, error = _error)
-    except Exception as e:
-      await _cleanup_directory(working_path)
-      return VectorDbResponse(success = False, error = f"Failed to download file {e}")
-
-  if rq.endpoints:
-    try:
-      documents.extend(await load_endpoints(rq.endpoints, working_path))
-    except Exception as e:
-      await _cleanup_directory(working_path)
-      tb = traceback.format_exc()
-      _error = f"Endpoint failure: {e} | {tb}"
-      logging.warning(_error)
-      return VectorDbResponse(success = False, error = _error)
-
-  if rq.integrations:
-    tasks = []
-    for integration in rq.integrations:
-      tasks.append(
-        IntegrationManager
-        .get_strategy(integration.type.lower(), working_path, integration.__pydantic_extra__) #type: ignore
-        .load()
-      )
-    try:
-      integration_documents = await asyncio.gather(*tasks)
-      for docs in integration_documents:
-        documents.extend(docs)
-    except Exception as e:
-      await _cleanup_directory(working_path)
-      tb = traceback.format_exc()
-      _error = f"Integration failure: {e} | {tb}"
-      logging.warning(_error)
-      return VectorDbResponse(success=False, error=_error)
-
-  #cleanup
-  await _cleanup_directory(working_path)
-
-  if documents and len(documents) > 0:
-    try:
-      store_path = os.path.join(_config.robot_data_folder, _config.robot_data_db_folder, _config.robot_data_db_folder_store, db_name)
-      db_file_path = await aiofiles.os.wrap(shutil.make_archive)(
-          os.path.join(_config.robot_data_folder, _config.robot_data_db_folder, _config.robot_data_db_folder_out, db_name),
-          "zip",
-          await VectorDbManager.get_strategy(rq.vector_type).create(rq.embeddings(), documents, store_path, return_folder_path=True)
-      )
-      return VectorDbResponse(file = os.path.basename(db_file_path), vector_type=rq.vector_type)
-    except Exception as e:
-      await _cleanup_directory(store_path)
-      return VectorDbResponse(success = False, error = str(e))
-    finally:
-      del documents
-      gc.collect()
-  else:
-    _error = "No documents found in the knowledgebase folder"
-    logging.warning(_error)
-    return VectorDbResponse(success = False, error = _error)
-
-async def kb_stream_file(filename: str):
-    file_path = os.path.join(config.robot_data_folder, config.robot_data_db_folder, config.robot_data_db_folder_out, filename)
-    if not os.path.isfile(file_path):
-        raise HTTPException(status_code=404, detail="File not found")
-    def iter_file():
-        with open(file_path, mode="rb") as file:
-            while chunk := file.read(1024*8):
-                yield chunk
-    return StreamingResponse(iter_file(), media_type="application/octet-stream", headers={"Content-Disposition": f"attachment; filename={filename}"})
+import os, gc, shutil, logging, traceback
+import asyncio, aiofiles, aiofiles.os
+from fastapi import HTTPException
+from fastapi.responses import StreamingResponse
+from langchain_core.documents import Document
+from ws_bom_robot_app.llm.vector_store.loader.base import Loader
+from ws_bom_robot_app.llm.models.api import RulesRequest, KbRequest, VectorDbResponse
+from ws_bom_robot_app.llm.vector_store.integration.manager import IntegrationManager
+from ws_bom_robot_app.llm.vector_store.db.manager import VectorDbManager
+from ws_bom_robot_app.config import config
+from ws_bom_robot_app.llm.models.kb import load_endpoints
+from ws_bom_robot_app.llm.utils.download import download_files
+
+async def _cleanup_directory(directory_path: str):
+  if os.path.exists(directory_path):
+    await asyncio.to_thread(shutil.rmtree, directory_path)
+
+#@timer
+async def rules(rq: RulesRequest) -> VectorDbResponse:
+  _config = rq.config()
+  db_name = rq.out_name()
+  store_path = os.path.join(_config.robot_data_folder, _config.robot_data_db_folder, _config.robot_data_db_folder_store, db_name)
+  try:
+    await VectorDbManager.get_strategy(rq.vector_type).create(rq.embeddings(),[Document(page_content=rule, metadata={"source": "rules"}) for rule in rq.rules], store_path) #type: ignore
+    db_file_path = shutil.make_archive(os.path.join(_config.robot_data_folder, _config.robot_data_db_folder, _config.robot_data_db_folder_out, db_name), "zip", store_path)
+    return VectorDbResponse(file = os.path.basename(db_file_path), vector_type=rq.vector_type)
+  except Exception as e:
+    try:
+      await _cleanup_directory(store_path)
+    finally:
+      return VectorDbResponse(success = False, error = str(e))
+  finally:
+    gc.collect()
+
+#@atimer
+async def kb(rq: KbRequest) -> VectorDbResponse:
+  os.environ['MPLCONFIGDIR'] = './tmp/.matplotlib'
+  _config = rq.config()
+  db_name = rq.out_name()
+  src_path = os.path.join(_config.robot_data_folder, _config.robot_data_db_folder, _config.robot_data_db_folder_src)
+  working_path = os.path.join(src_path, db_name)
+
+  if all([not rq.files,not rq.endpoints,not rq.integrations]):
+    return VectorDbResponse(success = False, error = "No files, endpoints or integrations provided")
+  else:
+    await aiofiles.os.makedirs(src_path, exist_ok=True)
+    await aiofiles.os.makedirs(working_path, exist_ok=True)
+
+  documents: list[Document] = []
+  # Download/copy all files
+  if rq.files:
+    try:
+      loaders = Loader(working_path)
+      filter_file_extensions = loaders.managed_file_extensions()
+      files_to_download = [file for file in rq.files if not os.path.exists(os.path.join(src_path, os.path.basename(file)))]
+      if files_to_download:
+        await download_files(
+          [f"{_config.robot_cms_host}/{_config.robot_cms_kb_folder}/{os.path.basename(file)}" for file in files_to_download if any([file.endswith(ext) for ext in filter_file_extensions])],
+          src_path, authorization=_config.robot_cms_auth)
+      # copy files to working tmp folder
+      for file in rq.files:
+        async with aiofiles.open(os.path.join(src_path, os.path.basename(file)), 'rb') as src_file:
+          async with aiofiles.open(os.path.join(working_path, os.path.basename(file)), 'wb') as dest_file:
+            await dest_file.write(await src_file.read())
+      #load files
+      try:
+        documents.extend(await loaders.load())
+      except Exception as e:
+        tb = traceback.format_exc()
+        _error = f"File loader failure: {e} | {tb}"
+        logging.warning(_error)
+        return VectorDbResponse(success = False, error = _error)
+    except Exception as e:
+      await _cleanup_directory(working_path)
+      return VectorDbResponse(success = False, error = f"Failed to download file {e}")
+
+  if rq.endpoints:
+    try:
+      documents.extend(await load_endpoints(rq.endpoints, working_path))
+    except Exception as e:
+      await _cleanup_directory(working_path)
+      tb = traceback.format_exc()
+      _error = f"Endpoint failure: {e} | {tb}"
+      logging.warning(_error)
+      return VectorDbResponse(success = False, error = _error)
+
+  if rq.integrations:
+    tasks = []
+    for integration in rq.integrations:
+      tasks.append(
+        IntegrationManager
+        .get_strategy(integration.type.lower(), working_path, integration.__pydantic_extra__) #type: ignore
+        .load()
+      )
+    try:
+      integration_documents = await asyncio.gather(*tasks)
+      for docs in integration_documents:
+        documents.extend(docs)
+    except Exception as e:
+      await _cleanup_directory(working_path)
+      tb = traceback.format_exc()
+      _error = f"Integration failure: {e} | {tb}"
+      logging.warning(_error)
+      return VectorDbResponse(success=False, error=_error)
+
+  #cleanup
+  await _cleanup_directory(working_path)
+
+  if documents and len(documents) > 0:
+    try:
+      store_path = os.path.join(_config.robot_data_folder, _config.robot_data_db_folder, _config.robot_data_db_folder_store, db_name)
+      db_file_path = await aiofiles.os.wrap(shutil.make_archive)(
+          os.path.join(_config.robot_data_folder, _config.robot_data_db_folder, _config.robot_data_db_folder_out, db_name),
+          "zip",
+          await VectorDbManager.get_strategy(rq.vector_type).create(rq.embeddings(), documents, store_path, return_folder_path=True)
+      )
+      return VectorDbResponse(file = os.path.basename(db_file_path), vector_type=rq.vector_type)
+    except Exception as e:
+      await _cleanup_directory(store_path)
+      return VectorDbResponse(success = False, error = str(e))
+    finally:
+      del documents
+      gc.collect()
+  else:
+    _error = "No documents found in the knowledgebase folder"
+    logging.warning(_error)
+    return VectorDbResponse(success = False, error = _error)
+
+async def kb_stream_file(filename: str):
+    file_path = os.path.join(config.robot_data_folder, config.robot_data_db_folder, config.robot_data_db_folder_out, filename)
+    if not os.path.isfile(file_path):
+        raise HTTPException(status_code=404, detail="File not found")
+    def iter_file():
+        with open(file_path, mode="rb") as file:
+            while chunk := file.read(1024*8):
+                yield chunk
+    return StreamingResponse(iter_file(), media_type="application/octet-stream", headers={"Content-Disposition": f"attachment; filename={filename}"})

ws_bom_robot_app/llm/vector_store/integration/manager.py

@@ -13,6 +13,7 @@ from ws_bom_robot_app.llm.vector_store.integration.sharepoint import Sharepoint
 from ws_bom_robot_app.llm.vector_store.integration.sitemap import Sitemap
 from ws_bom_robot_app.llm.vector_store.integration.slack import Slack
 from ws_bom_robot_app.llm.vector_store.integration.thron import Thron
+from ws_bom_robot_app.llm.vector_store.integration.shopify import Shopify
 class IntegrationManager:
   _list: dict[str, Type[IntegrationStrategy]] = {
     "llmkbazure": Azure,
@@ -28,6 +29,7 @@ class IntegrationManager:
     "llmkbsitemap": Sitemap,
     "llmkbslack": Slack,
     "llmkbthron": Thron,
+    "llmkbshopify": Shopify,
   }
   @classmethod
   def get_strategy(cls, name: str, knowledgebase_path: str, data: dict[str, str]) -> IntegrationStrategy:

ws_bom_robot_app/llm/vector_store/integration/shopify.py

@@ -0,0 +1,144 @@
+import asyncio, logging, aiohttp
+from ws_bom_robot_app.llm.vector_store.integration.base import IntegrationStrategy, UnstructuredIngest
+from unstructured_ingest.v2.processes.connectors.fsspec.sftp import SftpConnectionConfig, SftpAccessConfig, SftpDownloaderConfig, SftpIndexerConfig
+from langchain_core.documents import Document
+from ws_bom_robot_app.llm.vector_store.loader.base import Loader
+from typing import List, Union, Optional
+from pydantic import BaseModel, Field, AliasChoices, field_validator
+import json
+import os
+
+class ShopifyParams(BaseModel):
+  """
+  ShopifyParams is a model that defines the parameters required for Shopify integration.
+
+  Attributes:
+    shop_name (str): The shop name for Shopify.
+    access_token (str): The access token for Shopify.
+    graphql_query (Union[str, dict]): The GraphQL query string or dict for Shopify.
+  """
+  shop_name: str = Field(validation_alias=AliasChoices("shopName","shop_name"))
+  access_token: str = Field(validation_alias=AliasChoices("accessToken","access_token"))
+  graphql_query: Union[str, dict] = Field(validation_alias=AliasChoices("graphqlQuery","graphql_query"))
+
+  @field_validator('graphql_query')
+  @classmethod
+  def extract_query_string(cls, v):
+    """Extract the query string from dict format if needed"""
+    if isinstance(v, dict) and 'query' in v:
+      return v['query']
+    return v
+
+class Shopify(IntegrationStrategy):
+  def __init__(self, knowledgebase_path: str, data: dict[str, Union[str,int,list]]):
+    super().__init__(knowledgebase_path, data)
+    self.__data = ShopifyParams.model_validate(self.data)
+
+  def working_subdirectory(self) -> str:
+    return 'shopify'
+
+  async def run(self) -> None:
+    _data = await self.__get_data()
+    json_file_path = os.path.join(self.working_directory, 'shopify_data.json')
+    with open(json_file_path, 'w', encoding='utf-8') as f:
+      json.dump(_data, f, ensure_ascii=False)
+
+  async def load(self) -> list[Document]:
+    await self.run()
+    await asyncio.sleep(1)
+    return await Loader(self.working_directory).load()
+
+  async def __get_data(self, page_size: int = 50) -> List[dict]:
+      # URL dell'API
+    url = f"https://{self.__data.shop_name}.myshopify.com/admin/api/2024-07/graphql.json"
+
+    # Headers
+    headers = {
+        "X-Shopify-Access-Token": self.__data.access_token,
+        "Content-Type": "application/json"
+    }
+
+    all_products: List[dict] = []
+    has_next_page = True
+    cursor = None
+    retry_count = 0
+    max_retries = 5
+
+    while has_next_page:
+        # Variables per la query
+        variables = {
+            "first": page_size
+        }
+
+        if cursor:
+            variables["after"] = cursor
+
+        # Payload della richiesta
+        payload = {
+            "query": self.__data.graphql_query,
+            "variables": variables
+        }
+
+        try:
+            # Effettua la richiesta
+            async with aiohttp.ClientSession() as session:
+              async with session.post(url, headers=headers, json=payload) as response:
+                # Controlla se la risposta è JSON
+                try:
+                  data = await response.json()
+                except aiohttp.ContentTypeError:
+                  text = await response.text()
+                  logging.error(f"Non-JSON response received. Status code: {response.status}")
+                  logging.error(f"Content: {text}")
+                  raise Exception("Invalid response from API")
+
+            # Gestione del throttling
+            if "errors" in data:
+                error = data["errors"][0]
+                if error.get("extensions", {}).get("code") == "THROTTLED":
+                    retry_count += 1
+                    if retry_count > max_retries:
+                        raise Exception("Too many throttling attempts. Stopping execution.")
+
+                    # Aspetta un po' più a lungo ad ogni tentativo
+                    wait_time = 2 ** retry_count  # Backoff esponenziale
+                    print(f"Rate limit reached. Waiting {wait_time} seconds... (Attempt {retry_count}/{max_retries})")
+                    await asyncio.sleep(wait_time)
+                    continue
+                else:
+                    raise Exception(f"GraphQL errors: {data['errors']}")
+
+            # Resetta il contatore dei retry se la richiesta è andata bene
+            retry_count = 0
+
+            # Estrae i dati
+            products_data = data["data"]["products"]
+            edges = products_data["edges"]
+            page_info = products_data["pageInfo"]
+
+            # Aggiungi i prodotti alla lista
+            for edge in edges:
+                all_products.append(edge["node"])
+
+            # Aggiorna il cursore e il flag per la paginazione
+            has_next_page = page_info["hasNextPage"]
+            cursor = page_info["endCursor"]
+
+            print(f"Recuperati {len(edges)} prodotti. Totale: {len(all_products)}")
+
+            # Piccola pausa per evitare di saturare l'API
+            await asyncio.sleep(0.1)
+
+        except aiohttp.ClientError as e:
+            logging.error(f"Connection error: {e}")
+            retry_count += 1
+            if retry_count <= max_retries:
+                wait_time = 2 ** retry_count
+                logging.warning(f"Retrying in {wait_time} seconds...")
+                await asyncio.sleep(wait_time)
+                continue
+            else:
+                raise Exception("Too many network errors. Stopping execution.")
+
+    logging.info(f"Data retrieval completed! Total products: {len(all_products)}")
+    return all_products

ws_bom_robot_app/llm/vector_store/integration/thron.py

@@ -1,103 +1,103 @@
-import asyncio, logging, aiohttp
-from ws_bom_robot_app.llm.vector_store.integration.base import IntegrationStrategy, UnstructuredIngest
-from unstructured_ingest.v2.processes.connectors.fsspec.sftp import SftpConnectionConfig, SftpAccessConfig, SftpDownloaderConfig, SftpIndexerConfig
-from langchain_core.documents import Document
-from ws_bom_robot_app.llm.vector_store.loader.base import Loader
-from typing import List, Union, Optional
-from pydantic import BaseModel, Field, AliasChoices
-import json
-import os
-
-class ThronParams(BaseModel):
-  """
-  ThronParams is a model that defines the parameters required for Thron integration.
-
-  Attributes:
-    app_id (str): The application ID for Thron.
-    client_id (str): The client ID for Thron.
-    client_secret (str): The client secret for Thron.
-  """
-  organization_name: str = Field(validation_alias=AliasChoices("organizationName","organization_name"))
-  attribute_fields: Optional[List[str]] = Field(default=None, validation_alias=AliasChoices("attributeFields","attribute_fields"))
-  client_id: str = Field(validation_alias=AliasChoices("clientId","client_id"))
-  client_secret: str = Field(validation_alias=AliasChoices("clientSecret","client_secret"))
-
-class Thron(IntegrationStrategy):
-  def __init__(self, knowledgebase_path: str, data: dict[str, Union[str,int,list]]):
-    super().__init__(knowledgebase_path, data)
-    self.__data = ThronParams.model_validate(self.data)
-
-  def working_subdirectory(self) -> str:
-    return 'thron'
-
-  async def run(self) -> None:
-    _data = await self.__get_data()
-    transformed_data = self.__transform_data(_data)
-    json_file_path = os.path.join(self.working_directory, 'thron_data.json')
-    with open(json_file_path, 'w', encoding='utf-8') as f:
-      json.dump(transformed_data, f, indent=2, ensure_ascii=False)
-
-  async def load(self) -> list[Document]:
-    await self.run()
-    await asyncio.sleep(1)
-    return await Loader(self.working_directory).load()
-
-  async def __get_auth_token(self) -> str:
-    try:
-      async with aiohttp.ClientSession() as session:
-        auth_data = {
-          "grant_type": "client_credentials",
-          "client_id": self.__data.client_id,
-          "client_secret": self.__data.client_secret
-        }
-        headers = {
-          "accept": "application/json",
-          "Content-Type": "application/x-www-form-urlencoded"
-        }
-        async with session.post("https://websolute.thron.com/api/v1/authentication/oauth2/token", data=auth_data, headers=headers) as response:
-          result = await response.json()
-          return result.get("access_token", "")
-    except Exception as e:
-      logging.error(f"Error fetching Thron auth token: {e}")
-      return None
-
-  async def __get_data(self) -> dict:
-    try:
-      token = await self.__get_auth_token()
-      if not token:
-        logging.error("Failed to obtain Thron authentication token.")
-        return {}
-      attribute_fields = ",".join(self.__data.attribute_fields) if self.__data.attribute_fields else ""
-      async with aiohttp.ClientSession() as session:
-        headers = {
-          "accept": "application/json",
-          "Authorization": f"Bearer {token}"
-        }
-        async with session.get(f"https://{self.__data.organization_name}.thron.com/api/v1/product-data/products?attributeFields=product_id,{attribute_fields}", headers=headers) as response:
-          result = await response.json()
-          return result.get("items", {})
-    except Exception as e:
-      logging.error(f"Error fetching Thron product data: {e}")
-      return {}
-    return []
-
-
-
-  def __transform_data(self, data: dict) -> dict:
-    _data = []
-    for item in data:
-      if item.get("hierarchyLevel") == "MASTER":
-        # Iterate through variants to find the product_id
-        for item_variant in data:
-          if item_variant.get("hierarchyLevel") == "VARIANT":
-            for attr in item.get("attributes", []):
-              if attr.get("code") == "product_id" and attr.get("identifier") == item_variant.get("variation").get("master").split(":")[-1]:
-                # Initialize variants list if it doesn't exist
-                if "variants" not in item:
-                  item["variants"] = []
-                item["variants"].append(item_variant)
-                _data.append(item)
-                break
-      elif item.get("hierarchyLevel") == "SIMPLE":
-        _data.append(item)
-    return _data
+import asyncio, logging, aiohttp
+from ws_bom_robot_app.llm.vector_store.integration.base import IntegrationStrategy, UnstructuredIngest
+from unstructured_ingest.v2.processes.connectors.fsspec.sftp import SftpConnectionConfig, SftpAccessConfig, SftpDownloaderConfig, SftpIndexerConfig
+from langchain_core.documents import Document
+from ws_bom_robot_app.llm.vector_store.loader.base import Loader
+from typing import List, Union, Optional
+from pydantic import BaseModel, Field, AliasChoices
+import json
+import os
+
+class ThronParams(BaseModel):
+  """
+  ThronParams is a model that defines the parameters required for Thron integration.
+
+  Attributes:
+    app_id (str): The application ID for Thron.
+    client_id (str): The client ID for Thron.
+    client_secret (str): The client secret for Thron.
+  """
+  organization_name: str = Field(validation_alias=AliasChoices("organizationName","organization_name"))
+  attribute_fields: Optional[List[str]] = Field(default=None, validation_alias=AliasChoices("attributeFields","attribute_fields"))
+  client_id: str = Field(validation_alias=AliasChoices("clientId","client_id"))
+  client_secret: str = Field(validation_alias=AliasChoices("clientSecret","client_secret"))
+
+class Thron(IntegrationStrategy):
+  def __init__(self, knowledgebase_path: str, data: dict[str, Union[str,int,list]]):
+    super().__init__(knowledgebase_path, data)
+    self.__data = ThronParams.model_validate(self.data)
+
+  def working_subdirectory(self) -> str:
+    return 'thron'
+
+  async def run(self) -> None:
+    _data = await self.__get_data()
+    transformed_data = self.__transform_data(_data)
+    json_file_path = os.path.join(self.working_directory, 'thron_data.json')
+    with open(json_file_path, 'w', encoding='utf-8') as f:
+      json.dump(transformed_data, f, indent=2, ensure_ascii=False)
+
+  async def load(self) -> list[Document]:
+    await self.run()
+    await asyncio.sleep(1)
+    return await Loader(self.working_directory).load()
+
+  async def __get_auth_token(self) -> str:
+    try:
+      async with aiohttp.ClientSession() as session:
+        auth_data = {
+          "grant_type": "client_credentials",
+          "client_id": self.__data.client_id,
+          "client_secret": self.__data.client_secret
+        }
+        headers = {
+          "accept": "application/json",
+          "Content-Type": "application/x-www-form-urlencoded"
+        }
+        async with session.post("https://websolute.thron.com/api/v1/authentication/oauth2/token", data=auth_data, headers=headers) as response:
+          result = await response.json()
+          return result.get("access_token", "")
+    except Exception as e:
+      logging.error(f"Error fetching Thron auth token: {e}")
+      return None
+
+  async def __get_data(self) -> dict:
+    try:
+      token = await self.__get_auth_token()
+      if not token:
+        logging.error("Failed to obtain Thron authentication token.")
+        return {}
+      attribute_fields = ",".join(self.__data.attribute_fields) if self.__data.attribute_fields else ""
+      async with aiohttp.ClientSession() as session:
+        headers = {
+          "accept": "application/json",
+          "Authorization": f"Bearer {token}"
+        }
+        async with session.get(f"https://{self.__data.organization_name}.thron.com/api/v1/product-data/products?attributeFields=product_id,{attribute_fields}", headers=headers) as response:
+          result = await response.json()
+          return result.get("items", {})
+    except Exception as e:
+      logging.error(f"Error fetching Thron product data: {e}")
+      return {}
+    return []
+
+
+
+  def __transform_data(self, data: dict) -> dict:
+    _data = []
+    for item in data:
+      if item.get("hierarchyLevel") == "MASTER":
+        # Iterate through variants to find the product_id
+        for item_variant in data:
+          if item_variant.get("hierarchyLevel") == "VARIANT":
+            for attr in item.get("attributes", []):
+              if attr.get("code") == "product_id" and attr.get("identifier") == item_variant.get("variation").get("master").split(":")[-1]:
+                # Initialize variants list if it doesn't exist
+                if "variants" not in item:
+                  item["variants"] = []
+                item["variants"].append(item_variant)
+                _data.append(item)
+                break
+      elif item.get("hierarchyLevel") == "SIMPLE":
+        _data.append(item)
+    return _data