working es

application/parser/open_ai_func.py

@@ -1,8 +1,7 @@
 import os
 
 import tiktoken
-from langchain.embeddings import OpenAIEmbeddings
+from application.vectorstore.faiss import FaissStore
-from langchain.vectorstores import FAISS
 from retry import retry
 
 
@@ -33,11 +32,9 @@ def call_openai_api(docs, folder_name, task_status):
         os.makedirs(f"{folder_name}")
 
     from tqdm import tqdm
-    docs_test = [docs[0]]
-    docs.pop(0)
     c1 = 0
 
-    store = FAISS.from_documents(docs_test, OpenAIEmbeddings(openai_api_key=os.getenv("EMBEDDINGS_KEY")))
+    store = FaissStore(path=f"{folder_name}", embeddings_key=os.getenv("EMBEDDINGS_KEY"))
 
     # Uncomment for MPNet embeddings
     # model_name = "sentence-transformers/all-mpnet-base-v2"

application/requirements.txt

@@ -22,6 +22,7 @@ decorator==5.1.1
 dill==0.3.6
 dnspython==2.3.0
 ecdsa==0.18.0
+elasticsearch==8.9.0
 entrypoints==0.4
 faiss-cpu==1.7.3
 filelock==3.9.0

application/vectorstore/base.py

@@ -19,7 +19,7 @@ class BaseVectorStore(ABC):
     def is_azure_configured(self):
         return settings.OPENAI_API_BASE and settings.OPENAI_API_VERSION and settings.AZURE_DEPLOYMENT_NAME
 
-    def _get_docsearch(self, embeddings_name, embeddings_key=None):
+    def _get_embeddings(self, embeddings_name, embeddings_key=None):
         embeddings_factory = {
             "openai_text-embedding-ada-002": OpenAIEmbeddings,
             "huggingface_sentence-transformers/all-mpnet-base-v2": HuggingFaceHubEmbeddings,

application/vectorstore/elasticsearch.py

@@ -0,0 +1,199 @@
+from application.vectorstore.base import BaseVectorStore
+from application.core.settings import settings
+import elasticsearch
+#from langchain.vectorstores.elasticsearch import ElasticsearchStore
+
+
+class ElasticsearchStore(BaseVectorStore):
+    _es_connection = None  # Class attribute to hold the Elasticsearch connection
+
+    def __init__(self, path, embeddings_key, index_name="docsgpt"):
+        super().__init__()
+        self.path = path.replace("/app/application/indexes/", "")
+        self.embeddings_key = embeddings_key
+        self.index_name = index_name
+        
+        if ElasticsearchStore._es_connection is None:
+            connection_params = {}
+            connection_params["cloud_id"] = settings.ELASTIC_CLOUD_ID
+            connection_params["basic_auth"] = (settings.ELASTIC_USERNAME, settings.ELASTIC_PASSWORD)
+            ElasticsearchStore._es_connection = elasticsearch.Elasticsearch(**connection_params)
+            
+        self.docsearch = ElasticsearchStore._es_connection
+
+    def connect_to_elasticsearch(
+        *,
+        es_url = None,
+        cloud_id = None,
+        api_key = None,
+        username = None,
+        password = None,
+    ):
+        try:
+            import elasticsearch
+        except ImportError:
+            raise ImportError(
+                "Could not import elasticsearch python package. "
+                "Please install it with `pip install elasticsearch`."
+            )
+
+        if es_url and cloud_id:
+            raise ValueError(
+                "Both es_url and cloud_id are defined. Please provide only one."
+            )
+
+        connection_params = {}
+
+        if es_url:
+            connection_params["hosts"] = [es_url]
+        elif cloud_id:
+            connection_params["cloud_id"] = cloud_id
+        else:
+            raise ValueError("Please provide either elasticsearch_url or cloud_id.")
+
+        if api_key:
+            connection_params["api_key"] = api_key
+        elif username and password:
+            connection_params["basic_auth"] = (username, password)
+
+        es_client = elasticsearch.Elasticsearch(
+            **connection_params,
+        )
+        try:
+            es_client.info()
+        except Exception as e:
+            raise e
+
+        return es_client
+
+    def search(self, question, k=2, index_name=settings.ELASTIC_INDEX, *args, **kwargs):
+        embeddings = self._get_embeddings(settings.EMBEDDINGS_NAME, self.embeddings_key)
+        vector = embeddings.embed_query(question)
+        knn = {
+            "filter": [{"match": {"metadata.filename.keyword": self.path}}],
+            "field": "vector",
+            "k": k,
+            "num_candidates": 100,
+            "query_vector": vector,
+        }
+        full_query = {
+            "knn": knn,
+            "query": {
+                "bool": {
+                    "must": [
+                        {
+                            "match": {
+                                "text": {
+                                    "query": question,
+                                }
+                            }
+                        }
+                    ],
+                    "filter": [{"match": {"metadata.filename.keyword": self.path}}],
+                }
+            },
+            "rank": {"rrf": {}},
+        }
+        resp = self.docsearch.search(index=index_name, query=full_query['query'], size=k, knn=full_query['knn'])
+        return resp
+
+        def _create_index_if_not_exists(
+                self, index_name, dims_length
+            ):
+
+            if self.client.indices.exists(index=index_name):
+                print(f"Index {index_name} already exists.")
+
+            else:
+                self.strategy.before_index_setup(
+                    client=self.client,
+                    text_field=self.query_field,
+                    vector_query_field=self.vector_query_field,
+                )
+
+                indexSettings = self.index(
+                    dims_length=dims_length,
+                )
+                self.client.indices.create(index=index_name, **indexSettings)
+        def index(
+                self,
+                dims_length,
+            ):
+
+
+            return {
+                "mappings": {
+                    "properties": {
+                        "vector": {
+                            "type": "dense_vector",
+                            "dims": dims_length,
+                            "index": True,
+                            "similarity": "cosine",
+                        },
+                    }
+                }
+            }
+
+        def add_texts(
+            self,
+            texts,
+            metadatas = None,
+            ids = None,
+            refresh_indices = True,
+            create_index_if_not_exists = True,
+            bulk_kwargs = None,
+            **kwargs,
+        ):
+            
+            from elasticsearch.helpers import BulkIndexError, bulk
+
+            bulk_kwargs = bulk_kwargs or {}
+            import uuid
+            embeddings = []
+            ids = ids or [str(uuid.uuid4()) for _ in texts]
+            requests = []
+            embeddings = self._get_embeddings(settings.EMBEDDINGS_NAME, self.embeddings_key)
+
+            vectors = embeddings.embed_documents(list(texts))
+
+            dims_length = len(vectors[0])
+
+            if create_index_if_not_exists:
+                self._create_index_if_not_exists(
+                    index_name=self.index_name, dims_length=dims_length
+                )
+
+            for i, (text, vector) in enumerate(zip(texts, vectors)):
+                metadata = metadatas[i] if metadatas else {}
+
+                requests.append(
+                    {
+                        "_op_type": "index",
+                        "_index": self.index_name,
+                        "text": text,
+                        "vector": vector,
+                        "metadata": metadata,
+                        "_id": ids[i],
+                    }
+                )
+
+
+            if len(requests) > 0:
+                try:
+                    success, failed = bulk(
+                        self.client,
+                        requests,
+                        stats_only=True,
+                        refresh=refresh_indices,
+                        **bulk_kwargs,
+                    )
+                    return ids
+                except BulkIndexError as e:
+                    print(f"Error adding texts: {e}")
+                    firstError = e.errors[0].get("index", {}).get("error", {})
+                    print(f"First error reason: {firstError.get('reason')}")
+                    raise e
+
+            else:
+                return []
+

application/vectorstore/faiss.py

@@ -8,8 +8,11 @@ class FaissStore(BaseVectorStore):
         super().__init__()
         self.path = path
         self.docsearch = FAISS.load_local(
-            self.path, self._get_docsearch(settings.EMBEDDINGS_NAME, settings.EMBEDDINGS_KEY)
+            self.path, self._get_embeddings(settings.EMBEDDINGS_NAME, settings.EMBEDDINGS_KEY)
         )
 
     def search(self, *args, **kwargs):
         return self.docsearch.similarity_search(*args, **kwargs)
+
+    def add_texts(self, *args, **kwargs):
+        return self.docsearch.add_texts(*args, **kwargs)

frontend/package-lock.json

@@ -1346,9 +1346,9 @@
       }
     },
     "node_modules/@typescript-eslint/eslint-plugin/node_modules/semver": {
-      "version": "7.3.8",
+      "version": "7.5.4",
-      "resolved": "https://registry.npmjs.org/semver/-/semver-7.3.8.tgz",
+      "resolved": "https://registry.npmjs.org/semver/-/semver-7.5.4.tgz",
-      "integrity": "sha512-NB1ctGL5rlHrPJtFDVIVzTyQylMLu9N9VICA6HSFJo8MCGVTMW6gfpicwKmmK/dAjTOrqu5l63JJOpDSrAis3A==",
+      "integrity": "sha512-1bCSESV6Pv+i21Hvpxp3Dx+pSD8lIPt8uVjRrxAUt/nbswYc+tK6Y2btiULjd4+fnq15PX+nqQDC7Oft7WkwcA==",
       "dev": true,
       "dependencies": {
         "lru-cache": "^6.0.0"
@@ -1490,9 +1490,9 @@
       }
     },
     "node_modules/@typescript-eslint/typescript-estree/node_modules/semver": {
-      "version": "7.3.8",
+      "version": "7.5.4",
-      "resolved": "https://registry.npmjs.org/semver/-/semver-7.3.8.tgz",
+      "resolved": "https://registry.npmjs.org/semver/-/semver-7.5.4.tgz",
-      "integrity": "sha512-NB1ctGL5rlHrPJtFDVIVzTyQylMLu9N9VICA6HSFJo8MCGVTMW6gfpicwKmmK/dAjTOrqu5l63JJOpDSrAis3A==",
+      "integrity": "sha512-1bCSESV6Pv+i21Hvpxp3Dx+pSD8lIPt8uVjRrxAUt/nbswYc+tK6Y2btiULjd4+fnq15PX+nqQDC7Oft7WkwcA==",
       "dev": true,
       "dependencies": {
         "lru-cache": "^6.0.0"
@@ -1549,9 +1549,9 @@
       }
     },
     "node_modules/@typescript-eslint/utils/node_modules/semver": {
-      "version": "7.3.8",
+      "version": "7.5.4",
-      "resolved": "https://registry.npmjs.org/semver/-/semver-7.3.8.tgz",
+      "resolved": "https://registry.npmjs.org/semver/-/semver-7.5.4.tgz",
-      "integrity": "sha512-NB1ctGL5rlHrPJtFDVIVzTyQylMLu9N9VICA6HSFJo8MCGVTMW6gfpicwKmmK/dAjTOrqu5l63JJOpDSrAis3A==",
+      "integrity": "sha512-1bCSESV6Pv+i21Hvpxp3Dx+pSD8lIPt8uVjRrxAUt/nbswYc+tK6Y2btiULjd4+fnq15PX+nqQDC7Oft7WkwcA==",
       "dev": true,
       "dependencies": {
         "lru-cache": "^6.0.0"
@@ -1991,9 +1991,9 @@
       }
     },
     "node_modules/builtins/node_modules/semver": {
-      "version": "7.3.8",
+      "version": "7.5.4",
-      "resolved": "https://registry.npmjs.org/semver/-/semver-7.3.8.tgz",
+      "resolved": "https://registry.npmjs.org/semver/-/semver-7.5.4.tgz",
-      "integrity": "sha512-NB1ctGL5rlHrPJtFDVIVzTyQylMLu9N9VICA6HSFJo8MCGVTMW6gfpicwKmmK/dAjTOrqu5l63JJOpDSrAis3A==",
+      "integrity": "sha512-1bCSESV6Pv+i21Hvpxp3Dx+pSD8lIPt8uVjRrxAUt/nbswYc+tK6Y2btiULjd4+fnq15PX+nqQDC7Oft7WkwcA==",
       "dev": true,
       "dependencies": {
         "lru-cache": "^6.0.0"
@@ -2055,9 +2055,9 @@
       }
     },
     "node_modules/caniuse-lite": {
-      "version": "1.0.30001450",
+      "version": "1.0.30001541",
-      "resolved": "https://registry.npmjs.org/caniuse-lite/-/caniuse-lite-1.0.30001450.tgz",
+      "resolved": "https://registry.npmjs.org/caniuse-lite/-/caniuse-lite-1.0.30001541.tgz",
-      "integrity": "sha512-qMBmvmQmFXaSxexkjjfMvD5rnDL0+m+dUMZKoDYsGG8iZN29RuYh9eRoMvKsT6uMAWlyUUGDEQGJJYjzCIO9ew==",
+      "integrity": "sha512-bLOsqxDgTqUBkzxbNlSBt8annkDpQB9NdzdTbO2ooJ+eC/IQcvDspDc058g84ejCelF7vHUx57KIOjEecOHXaw==",
       "dev": true,
       "funding": [
         {
@@ -2067,6 +2067,10 @@
         {
           "type": "tidelift",
           "url": "https://tidelift.com/funding/github/npm/caniuse-lite"
+        },
+        {
+          "type": "github",
+          "url": "https://github.com/sponsors/ai"
         }
       ]
     },
@@ -2889,9 +2893,9 @@
       }
     },
     "node_modules/eslint-plugin-n/node_modules/semver": {
-      "version": "7.3.8",
+      "version": "7.5.4",
-      "resolved": "https://registry.npmjs.org/semver/-/semver-7.3.8.tgz",
+      "resolved": "https://registry.npmjs.org/semver/-/semver-7.5.4.tgz",
-      "integrity": "sha512-NB1ctGL5rlHrPJtFDVIVzTyQylMLu9N9VICA6HSFJo8MCGVTMW6gfpicwKmmK/dAjTOrqu5l63JJOpDSrAis3A==",
+      "integrity": "sha512-1bCSESV6Pv+i21Hvpxp3Dx+pSD8lIPt8uVjRrxAUt/nbswYc+tK6Y2btiULjd4+fnq15PX+nqQDC7Oft7WkwcA==",
       "dev": true,
       "dependencies": {
         "lru-cache": "^6.0.0"
@@ -4478,9 +4482,9 @@
       }
     },
     "node_modules/lint-staged/node_modules/yaml": {
-      "version": "2.2.1",
+      "version": "2.3.2",
-      "resolved": "https://registry.npmjs.org/yaml/-/yaml-2.2.1.tgz",
+      "resolved": "https://registry.npmjs.org/yaml/-/yaml-2.3.2.tgz",
-      "integrity": "sha512-e0WHiYql7+9wr4cWMx3TVQrNwejKaEe7/rHNmQmqRjazfOP5W8PB6Jpebb5o6fIapbz9o9+2ipcaTM2ZwDI6lw==",
+      "integrity": "sha512-N/lyzTPaJasoDmfV7YTrYCI0G/3ivm/9wdG0aHuheKowWQwGTsK0Eoiw6utmzAnI6pkJa0DUVygvp3spqqEKXg==",
       "dev": true,
       "engines": {
         "node": ">= 14"
@@ -6532,9 +6536,9 @@
       }
     },
     "node_modules/semver": {
-      "version": "6.3.0",
+      "version": "6.3.1",
-      "resolved": "https://registry.npmjs.org/semver/-/semver-6.3.0.tgz",
+      "resolved": "https://registry.npmjs.org/semver/-/semver-6.3.1.tgz",
-      "integrity": "sha512-b39TBaTSfV6yBrapU89p5fKekE2m/NwnDocOVruQFS1/veMgdzuPcnOM34M6CwxW8jH/lxEa5rBoDeUwu5HHTw==",
+      "integrity": "sha512-BR7VvDCVHO+q2xBEWskxS6DJE1qRnb7DxzUrogb71CWoSficBxYsiAGd+Kl0mmq/MprG9yArRkyrQxTO6XjMzA==",
       "dev": true,
       "bin": {
         "semver": "bin/semver.js"