follow ups on RAG graph part

examples/LynxScribe Image RAG

@@ -68,7 +68,7 @@
           "type": "basic"
         },
         "params": {
-          "chat": "show me a picture about 2 doctors"
+          "chat": "show me a picture about cyclists"
         },
         "status": "done",
         "title": "Input chat"
@@ -304,9 +304,9 @@
         },
         "params": {
           "text_embedder_interface": "openai",
-          "text_embedder_model_name_or_path": "text-embedding-3-small",
+          "text_embedder_model_name_or_path": "text-embedding-3-large",
           "vdb_collection_name": "lynx",
-          "vdb_num_dimensions": "1536",
+          "vdb_num_dimensions": "3072",
           "vdb_provider_name": "faiss"
         },
         "status": "done",
@@ -388,7 +388,7 @@
     },
     {
       "data": {
-        "display": "https://storage.googleapis.com/lynxkite_public_data/lynxscribe-images/image-rag-test/surgery-1807541_1280.jpg",
+        "display": "https://storage.googleapis.com/lynxkite_public_data/lynxscribe-images/image-rag-test/riding-7661658_1280.jpg",
         "error": null,
         "meta": {
           "inputs": {
@@ -414,11 +414,11 @@
         "title": "LynxScribe Image Result Viewer"
       },
       "dragHandle": ".bg-primary",
-      "height": 622.0,
+      "height": 574.0,
       "id": "LynxScribe Image Result Viewer 1",
       "position": {
-        "x": 1550.5086064306404,
-        "y": -349.93521115271193
+        "x": 1551.9130513063783,
+        "y": -326.059648265166
       },
       "type": "image",
       "width": 802.0

examples/LynxScribe demo

@@ -50,25 +50,25 @@
       "targetHandle": "chat_processor"
     },
     {
-      "id": "Cloud-sourced File Loader 1 LynxScribe RAG Graph Chatbot Builder 1",
-      "source": "Cloud-sourced File Loader 1",
+      "id": "Cloud-sourced File Listing 1 LynxScribe Text RAG Loader 1",
+      "source": "Cloud-sourced File Listing 1",
       "sourceHandle": "output",
-      "target": "LynxScribe RAG Graph Chatbot Builder 1",
+      "target": "LynxScribe Text RAG Loader 1",
       "targetHandle": "file_urls"
     },
+    {
+      "id": "LynxScribe Text RAG Loader 1 LynxScribe RAG Graph Chatbot Builder 1",
+      "source": "LynxScribe Text RAG Loader 1",
+      "sourceHandle": "output",
+      "target": "LynxScribe RAG Graph Chatbot Builder 1",
+      "targetHandle": "rag_graph"
+    },
     {
       "id": "LynxScribe RAG Graph Chatbot Builder 1 LynxScribe RAG Graph Chatbot Backend 1",
       "source": "LynxScribe RAG Graph Chatbot Builder 1",
       "sourceHandle": "output",
       "target": "LynxScribe RAG Graph Chatbot Backend 1",
       "targetHandle": "knowledge_base"
-    },
-    {
-      "id": "LynxScribe RAG Graph Vector Store 1 LynxScribe RAG Graph Chatbot Builder 1",
-      "source": "LynxScribe RAG Graph Vector Store 1",
-      "sourceHandle": "output",
-      "target": "LynxScribe RAG Graph Chatbot Builder 1",
-      "targetHandle": "rag_graph"
     }
   ],
   "env": "LynxScribe",
@@ -103,7 +103,7 @@
           "type": "basic"
         },
         "params": {
-          "chat": "What products does Lynx have?"
+          "chat": "Wgo is Gabor?"
         },
         "status": "done",
         "title": "Input chat"
@@ -129,7 +129,7 @@
               ],
               "data": [
                 [
-                  "Lynx Analytics offers a range of data analytics products and solutions tailored for various industries. Here are some of our key offerings:\n\n- **Generative AI**: We provide innovative solutions such as chatbots specifically designed for the pharmaceutical and service provider sectors.\n  \n- **Graph AI**: Our advanced graph reasoning tools help in areas like transport scenario planning and predicting patient outcomes with graph representation learning.\n\n- **Pharma and Life Sciences**: We focus on marketing support, including Next Best Action predictions and Brand Adoption Ladder analysis, as well as supporting drug discovery and medical analytics.\n\n- **Retail Solutions**: Our products include Price AI, Assort AI, and Promo AI to optimize pricing, assortment, and promotion strategies for retailers.\n\n- **Financial Services**: We offer digital banking analytics solutions and a Customer Happiness Index to enhance customer experience and retention.\n\n- **Telecommunications**: While briefly mentioned, our telecom solutions help optimize operations, including fibre CAPEX optimization and churn prevention management.\n\nThese products are complemented by consulting services to help businesses make data-driven decisions. If you need more specific information or a demo, feel free to get in touch.\n\nCould you share what prompted your visit to our website today? Or may I know which specific domain or industry you are interested in or work in?"
+                  "Lynx Analytics has two notable professionals named G\u00e1bor. Could you please specify which G\u00e1bor you are inquiring about?\n\n- **G\u00e1bor Benedek**: Chief Innovation Officer & Co-founder at Lynx Analytics. He specializes in economic and business simulations, social network analysis, data mining, and predictive analytics. He has an academic background as a former Associate Professor at Corvinus University of Budapest and has founded several data-related companies.\n\n- **G\u00e1bor Kriv\u00e1chy**: Country Manager at Lynx Analytics in Hungary. He is an experienced technology executive with a background in system implementation, integration, and project management, particularly in SAP implementations.\n\nLet me know which G\u00e1bor's details you would like to learn more about!"
                 ]
               ]
             }
@@ -256,6 +256,8 @@
     },
     {
       "data": {
+        "__execution_delay": 0.0,
+        "collapsed": null,
         "display": null,
         "error": null,
         "meta": {
@@ -324,6 +326,8 @@
     },
     {
       "data": {
+        "__execution_delay": 0.0,
+        "collapsed": null,
         "display": null,
         "error": null,
         "meta": {
@@ -456,8 +460,23 @@
         "display": null,
         "error": null,
         "meta": {
-          "inputs": {},
-          "name": "LynxScribe RAG Graph Vector Store",
+          "inputs": {
+            "chat_processor": {
+              "name": "chat_processor",
+              "position": "bottom",
+              "type": {
+                "type": "<class 'inspect._empty'>"
+              }
+            },
+            "knowledge_base": {
+              "name": "knowledge_base",
+              "position": "bottom",
+              "type": {
+                "type": "<class 'inspect._empty'>"
+              }
+            }
+          },
+          "name": "LynxScribe RAG Graph Chatbot Backend",
           "outputs": {
             "output": {
               "name": "output",
@@ -468,67 +487,91 @@
             }
           },
           "params": {
-            "collection_name": {
-              "default": "lynx",
-              "name": "collection_name",
+            "llm_interface": {
+              "default": "openai",
+              "name": "llm_interface",
               "type": {
                 "type": "<class 'str'>"
               }
             },
-            "name": {
-              "default": "faiss",
-              "name": "name",
+            "llm_model_name": {
+              "default": "gpt-4o",
+              "name": "llm_model_name",
               "type": {
                 "type": "<class 'str'>"
               }
             },
-            "num_dimensions": {
-              "default": 3072.0,
-              "name": "num_dimensions",
+            "negative_answer": {
+              "default": "I'm sorry, but the data I've been trained on does not contain any information related to your question.",
+              "name": "negative_answer",
               "type": {
-                "type": "<class 'int'>"
+                "type": "<class 'str'>"
               }
             },
-            "text_embedder_interface": {
-              "default": "openai",
-              "name": "text_embedder_interface",
+            "retriever_limits_by_type": {
+              "default": "{}",
+              "name": "retriever_limits_by_type",
               "type": {
                 "type": "<class 'str'>"
               }
             },
-            "text_embedder_model_name_or_path": {
-              "default": "text-embedding-3-large",
-              "name": "text_embedder_model_name_or_path",
+            "retriever_max_iterations": {
+              "default": 3.0,
+              "name": "retriever_max_iterations",
               "type": {
-                "type": "<class 'str'>"
+                "type": "<class 'int'>"
+              }
+            },
+            "retriever_overall_chunk_limit": {
+              "default": 20.0,
+              "name": "retriever_overall_chunk_limit",
+              "type": {
+                "type": "<class 'int'>"
+              }
+            },
+            "retriever_overall_token_limit": {
+              "default": 3000.0,
+              "name": "retriever_overall_token_limit",
+              "type": {
+                "type": "<class 'int'>"
+              }
+            },
+            "retriever_strict_limits": {
+              "default": true,
+              "name": "retriever_strict_limits",
+              "type": {
+                "type": "<class 'bool'>"
               }
             }
           },
           "position": {
-            "x": 189.0,
-            "y": 489.0
+            "x": 543.0,
+            "y": 256.0
           },
           "type": "basic"
         },
         "params": {
-          "collection_name": "lynx",
-          "name": "faiss",
-          "num_dimensions": "1536",
-          "text_embedder_interface": "openai",
-          "text_embedder_model_name_or_path": "text-embedding-ada-002"
+          "llm_interface": "openai",
+          "llm_model_name": "gpt-4o",
+          "negative_answer": "I'm sorry, but the data I've been trained on does not contain any information related to your question.",
+          "retriever_limits_by_type": "{\"information\": [1, 5], \"summary\": [0, 2], \"template_qna\": [1, 3], \"QnA question\": [0, 0]}",
+          "retriever_max_iterations": 3.0,
+          "retriever_overall_chunk_limit": "20",
+          "retriever_overall_token_limit": 3000.0,
+          "retriever_strict_limits": true
         },
         "status": "done",
-        "title": "LynxScribe RAG Graph Vector Store"
+        "title": "LynxScribe RAG Graph Chatbot Backend"
       },
       "dragHandle": ".bg-primary",
-      "height": 443.0,
-      "id": "LynxScribe RAG Graph Vector Store 1",
+      "height": 556.0,
+      "id": "LynxScribe RAG Graph Chatbot Backend 1",
       "position": {
-        "x": -2019.279230344727,
-        "y": 1080.1955856484205
+        "x": -2020.0,
+        "y": -188.33333333333334
       },
       "type": "basic",
-      "width": 336.0
+      "width": 903.0
     },
     {
       "data": {
@@ -537,8 +580,16 @@
         "display": null,
         "error": null,
         "meta": {
-          "inputs": {},
-          "name": "Cloud-sourced File Loader",
+          "inputs": {
+            "file_urls": {
+              "name": "file_urls",
+              "position": "left",
+              "type": {
+                "type": "<class 'inspect._empty'>"
+              }
+            }
+          },
+          "name": "LynxScribe Text RAG Loader",
           "outputs": {
             "output": {
               "name": "output",
@@ -549,51 +600,78 @@
             }
           },
           "params": {
-            "accepted_file_types": {
-              "default": ".jpg, .jpeg, .png",
-              "name": "accepted_file_types",
+            "input_type": {
+              "default": "v1",
+              "name": "input_type",
+              "type": {
+                "enum": [
+                  "V1",
+                  "V2"
+                ]
+              }
+            },
+            "text_embedder_interface": {
+              "default": "openai",
+              "name": "text_embedder_interface",
               "type": {
                 "type": "<class 'str'>"
               }
             },
-            "cloud_provider": {
-              "default": "gcp",
-              "name": "cloud_provider",
+            "text_embedder_model_name_or_path": {
+              "default": "text-embedding-3-large",
+              "name": "text_embedder_model_name_or_path",
               "type": {
                 "type": "<class 'str'>"
               }
             },
-            "folder_URL": {
-              "default": "https://storage.googleapis.com/lynxkite_public_data/lynxscribe-images/image-rag-test",
-              "name": "folder_URL",
+            "vdb_collection_name": {
+              "default": "lynx",
+              "name": "vdb_collection_name",
+              "type": {
+                "type": "<class 'str'>"
+              }
+            },
+            "vdb_num_dimensions": {
+              "default": 3072.0,
+              "name": "vdb_num_dimensions",
+              "type": {
+                "type": "<class 'int'>"
+              }
+            },
+            "vdb_provider_name": {
+              "default": "faiss",
+              "name": "vdb_provider_name",
               "type": {
                 "type": "<class 'str'>"
               }
             }
           },
           "position": {
-            "x": 189.0,
-            "y": 412.0
+            "x": 870.0,
+            "y": 926.0
           },
           "type": "basic"
         },
         "params": {
-          "accepted_file_types": ".pickle",
-          "cloud_provider": "gcp",
-          "folder_URL": "https://storage.googleapis.com/lynxkite_public_data/lynxscribe-knowledge-graphs/lynx-chatbot"
+          "input_type": "V1",
+          "text_embedder_interface": "openai",
+          "text_embedder_model_name_or_path": "text-embedding-ada-002",
+          "vdb_collection_name": "lynx",
+          "vdb_num_dimensions": "1536",
+          "vdb_provider_name": "faiss"
         },
         "status": "done",
-        "title": "Cloud-sourced File Loader"
+        "title": "LynxScribe Text RAG Loader"
       },
       "dragHandle": ".bg-primary",
-      "height": 325.0,
-      "id": "Cloud-sourced File Loader 1",
+      "height": 520.0,
+      "id": "LynxScribe Text RAG Loader 1",
       "position": {
-        "x": -2884.195823214815,
-        "y": 630.6408986626046
+        "x": -2980.4063452955706,
+        "y": 787.1039827859594
       },
       "type": "basic",
-      "width": 622.0
+      "width": 318.0
     },
     {
       "data": {
@@ -602,118 +680,67 @@
         "display": null,
         "error": null,
         "meta": {
-          "inputs": {
-            "chat_processor": {
-              "name": "chat_processor",
-              "position": "bottom",
-              "type": {
-                "type": "<class 'inspect._empty'>"
-              }
-            },
-            "knowledge_base": {
-              "name": "knowledge_base",
-              "position": "bottom",
-              "type": {
-                "type": "<class 'inspect._empty'>"
-              }
-            }
-          },
-          "name": "LynxScribe RAG Graph Chatbot Backend",
+          "inputs": {},
+          "name": "Cloud-sourced File Listing",
           "outputs": {
             "output": {
               "name": "output",
-              "position": "top",
+              "position": "right",
               "type": {
                 "type": "None"
               }
             }
           },
           "params": {
-            "llm_interface": {
-              "default": "openai",
-              "name": "llm_interface",
-              "type": {
-                "type": "<class 'str'>"
-              }
-            },
-            "llm_model_name": {
-              "default": "gpt-4o",
-              "name": "llm_model_name",
+            "accepted_file_types": {
+              "default": ".jpg, .jpeg, .png",
+              "name": "accepted_file_types",
               "type": {
                 "type": "<class 'str'>"
               }
             },
-            "negative_answer": {
-              "default": "I'm sorry, but the data I've been trained on does not contain any information related to your question.",
-              "name": "negative_answer",
+            "cloud_provider": {
+              "default": "gcp",
+              "name": "cloud_provider",
               "type": {
-                "type": "<class 'str'>"
+                "enum": [
+                  "GCP",
+                  "AWS",
+                  "AZURE"
+                ]
               }
             },
-            "retriever_limits_by_type": {
-              "default": "{}",
-              "name": "retriever_limits_by_type",
+            "folder_URL": {
+              "default": "https://storage.googleapis.com/lynxkite_public_data/lynxscribe-images/image-rag-test",
+              "name": "folder_URL",
               "type": {
                 "type": "<class 'str'>"
               }
-            },
-            "retriever_max_iterations": {
-              "default": 3.0,
-              "name": "retriever_max_iterations",
-              "type": {
-                "type": "<class 'int'>"
-              }
-            },
-            "retriever_overall_chunk_limit": {
-              "default": 20.0,
-              "name": "retriever_overall_chunk_limit",
-              "type": {
-                "type": "<class 'int'>"
-              }
-            },
-            "retriever_overall_token_limit": {
-              "default": 3000.0,
-              "name": "retriever_overall_token_limit",
-              "type": {
-                "type": "<class 'int'>"
-              }
-            },
-            "retriever_strict_limits": {
-              "default": true,
-              "name": "retriever_strict_limits",
-              "type": {
-                "type": "<class 'bool'>"
-              }
             }
           },
           "position": {
-            "x": 543.0,
-            "y": 256.0
+            "x": 451.0,
+            "y": 505.0
           },
           "type": "basic"
         },
         "params": {
-          "llm_interface": "openai",
-          "llm_model_name": "gpt-4o",
-          "negative_answer": "I'm sorry, but the data I've been trained on does not contain any information related to your question.",
-          "retriever_limits_by_type": "{\"information\": [1, 5], \"summary\": [0, 2], \"template_qna\": [1, 3], \"QnA question\": [0, 0]}",
-          "retriever_max_iterations": 3.0,
-          "retriever_overall_chunk_limit": "20",
-          "retriever_overall_token_limit": 3000.0,
-          "retriever_strict_limits": true
+          "accepted_file_types": ".pickle",
+          "cloud_provider": "GCP",
+          "folder_URL": "https://storage.googleapis.com/lynxkite_public_data/lynxscribe-knowledge-graphs/lynx-chatbot"
         },
         "status": "done",
-        "title": "LynxScribe RAG Graph Chatbot Backend"
+        "title": "Cloud-sourced File Listing"
       },
       "dragHandle": ".bg-primary",
-      "height": 556.0,
-      "id": "LynxScribe RAG Graph Chatbot Backend 1",
+      "height": 324.0,
+      "id": "Cloud-sourced File Listing 1",
       "position": {
-        "x": -2020.0,
-        "y": -188.33333333333334
+        "x": -3827.1644268005352,
+        "y": 883.7859821532916
       },
       "type": "basic",
-      "width": 903.0
+      "width": 613.0
     },
     {
       "data": {
@@ -721,16 +748,9 @@
         "error": null,
         "meta": {
           "inputs": {
-            "file_urls": {
-              "name": "file_urls",
-              "position": "left",
-              "type": {
-                "type": "<class 'inspect._empty'>"
-              }
-            },
             "rag_graph": {
               "name": "rag_graph",
-              "position": "bottom",
+              "position": "left",
               "type": {
                 "type": "<class 'inspect._empty'>"
               }
@@ -747,13 +767,6 @@
             }
           },
           "params": {
-            "input_type": {
-              "default": "v1",
-              "name": "input_type",
-              "type": {
-                "type": "<class 'str'>"
-              }
-            },
             "node_types": {
               "default": "intent_cluster",
               "name": "node_types",
@@ -770,13 +783,12 @@
             }
           },
           "position": {
-            "x": 466.0,
-            "y": 478.0
+            "x": 1314.0,
+            "y": 1003.0
           },
           "type": "basic"
         },
         "params": {
-          "input_type": "v1",
           "node_types": "intent_cluster",
           "scenario_file": "uploads/lynx_chatbot_scenario_selector.yaml"
         },
@@ -784,14 +796,14 @@
         "title": "LynxScribe RAG Graph Chatbot Builder"
       },
       "dragHandle": ".bg-primary",
-      "height": 364.0,
+      "height": 238.0,
       "id": "LynxScribe RAG Graph Chatbot Builder 1",
       "position": {
-        "x": -2101.666666666667,
-        "y": 611.6666666666666
+        "x": -2453.755433834285,
+        "y": 927.5600547745715
       },
       "type": "basic",
-      "width": 500.0
+      "width": 448.0
     }
   ]
 }

lynxkite-lynxscribe/src/lynxkite_lynxscribe/lynxscribe_ops.py

@@ -34,11 +34,6 @@ from lynxkite.core.executors import one_by_one
 
 DEFAULT_NEGATIVE_ANSWER = "I'm sorry, but the data I've been trained on does not contain any information related to your question."
 
-# logger
-# import logging
-# logging.basicConfig(level=logging.INFO)
-# logger = logging.getLogger(__name__)
-
 ENV = "LynxScribe"
 one_by_one.register(ENV)
 mem = joblib.Memory("joblib-cache")
@@ -53,6 +48,11 @@ class CloudProvider(Enum):
     AZURE = "azure"
 
 
+class RAGVersion(Enum):
+    V1 = "v1"
+    V2 = "v2"
+
+
 @op("Cloud-sourced File Listing")
 def cloud_file_loader(
     *,
@@ -88,43 +88,43 @@ def cloud_file_loader(
         raise ValueError(f"Cloud provider '{cloud_provider}' is not supported.")
 
 
-@output_on_top
-@op("LynxScribe RAG Graph Vector Store")
-@mem.cache
-def ls_rag_graph(
-    *,
-    name: str = "faiss",
-    num_dimensions: int = 3072,
-    collection_name: str = "lynx",
-    text_embedder_interface: str = "openai",
-    text_embedder_model_name_or_path: str = "text-embedding-3-large",
-    # api_key_name: str = "OPENAI_API_KEY",
-):
-    """
-    Returns with a vector store instance.
-    """
-
-    # getting the text embedder instance
-    llm_params = {"name": text_embedder_interface}
-    # if api_key_name:
-    #     llm_params["api_key"] = os.getenv(api_key_name)
-    llm = get_llm_engine(**llm_params)
-    text_embedder = TextEmbedder(llm=llm, model=text_embedder_model_name_or_path)
-
-    # getting the vector store
-    if name == "chromadb":
-        vector_store = get_vector_store(name=name, collection_name=collection_name)
-    elif name == "faiss":
-        vector_store = get_vector_store(name=name, num_dimensions=num_dimensions)
-    else:
-        raise ValueError(f"Vector store name '{name}' is not supported.")
-
-    # building up the RAG graph
-    rag_graph = RAGGraph(
-        PandasKnowledgeBaseGraph(vector_store=vector_store, text_embedder=text_embedder)
-    )
+# @output_on_top
+# @op("LynxScribe RAG Graph Vector Store")
+# @mem.cache
+# def ls_rag_graph(
+#     *,
+#     name: str = "faiss",
+#     num_dimensions: int = 3072,
+#     collection_name: str = "lynx",
+#     text_embedder_interface: str = "openai",
+#     text_embedder_model_name_or_path: str = "text-embedding-3-large",
+#     # api_key_name: str = "OPENAI_API_KEY",
+# ):
+#     """
+#     Returns with a vector store instance.
+#     """
+
+#     # getting the text embedder instance
+#     llm_params = {"name": text_embedder_interface}
+#     # if api_key_name:
+#     #     llm_params["api_key"] = os.getenv(api_key_name)
+#     llm = get_llm_engine(**llm_params)
+#     text_embedder = TextEmbedder(llm=llm, model=text_embedder_model_name_or_path)
+
+#     # getting the vector store
+#     if name == "chromadb":
+#         vector_store = get_vector_store(name=name, collection_name=collection_name)
+#     elif name == "faiss":
+#         vector_store = get_vector_store(name=name, num_dimensions=num_dimensions)
+#     else:
+#         raise ValueError(f"Vector store name '{name}' is not supported.")
+
+#     # building up the RAG graph
+#     rag_graph = RAGGraph(
+#         PandasKnowledgeBaseGraph(vector_store=vector_store, text_embedder=text_embedder)
+#     )
 
-    return {"rag_graph": rag_graph}
+#     return {"rag_graph": rag_graph}
 
 
 @op("LynxScribe Image Describer")
@@ -296,7 +296,7 @@ async def ls_image_rag_builder(
     # # saving the RAG graph
     # rag_graph.kg_base.save(image_rag_out_path)
 
-    return {"knowledge_base": rag_graph}
+    return {"rag_graph": rag_graph}
 
 
 @op("LynxScribe RAG Graph Saver")
@@ -320,7 +320,7 @@ def ls_save_rag_graph(
 @op("LynxScribe Image RAG Query")
 async def search_context(rag_graph, text, *, top_k=3):
     message = text["text"]
-    rag_graph = rag_graph[0]["knowledge_base"]
+    rag_graph = rag_graph[0]["rag_graph"]
 
     # get all similarities
     emb_similarities = await rag_graph.search_context(
@@ -352,90 +352,92 @@ def view_image(embedding_similarities):
     return embedding_similarities[0]["image_url"]
 
 
-# @output_on_top
-# @op("Vector store")
-# def vector_store(*, name="chromadb", collection_name="lynx"):
-#     vector_store = get_vector_store(name=name, collection_name=collection_name)
-#     return {"vector_store": vector_store}
+@op("LynxScribe Text RAG Loader")
+@mem.cache
+def ls_text_rag_loader(
+    file_urls,
+    *,
+    input_type: RAGVersion = RAGVersion.V1,
+    vdb_provider_name: str = "faiss",
+    vdb_num_dimensions: int = 3072,
+    vdb_collection_name: str = "lynx",
+    text_embedder_interface: str = "openai",
+    text_embedder_model_name_or_path: str = "text-embedding-3-large",
+    # api_key_name: str = "OPENAI_API_KEY",
+):
+    """
+    Loading a text-based RAG graph from saved files (getting pandas readable links).
+    """
 
+    # handling inputs
+    file_urls = file_urls["file_urls"]
 
-# @output_on_top
-# @op("LLM")
-# def llm(*, name="openai"):
-#     llm = get_llm_engine(name=name)
-#     return {"llm": llm}
+    # getting the text embedder instance
+    llm_params = {"name": text_embedder_interface}
+    # if api_key_name:
+    #     llm_params["api_key"] = os.getenv(api_key_name)
+    llm = get_llm_engine(**llm_params)
+    text_embedder = TextEmbedder(llm=llm, model=text_embedder_model_name_or_path)
 
+    # getting the vector store
+    if vdb_provider_name == "chromadb":
+        vector_store = get_vector_store(
+            name=vdb_provider_name, collection_name=vdb_collection_name
+        )
+    elif vdb_provider_name == "faiss":
+        vector_store = get_vector_store(
+            name=vdb_provider_name, num_dimensions=vdb_num_dimensions
+        )
+    else:
+        raise ValueError(f"Vector store name '{vdb_provider_name}' is not supported.")
 
-# @output_on_top
-# @ops.input_position(llm="bottom")
-# @op("Text embedder")
-# def text_embedder(llm, *, model="text-embedding-ada-002"):
-#     llm = llm[0]["llm"]
-#     text_embedder = TextEmbedder(llm=llm, model=model)
-#     return {"text_embedder": text_embedder}
+    # building up the RAG graph
+    rag_graph = RAGGraph(
+        PandasKnowledgeBaseGraph(vector_store=vector_store, text_embedder=text_embedder)
+    )
 
+    # loading the knowledge base (temporary + TODO: adding v2)
+    if input_type == RAGVersion.V1:
+        node_file = [f for f in file_urls if "nodes.p" in f][0]
+        edge_file = [f for f in file_urls if "edges.p" in f][0]
+        tempcluster_file = [f for f in file_urls if "clusters.p" in f][0]
+        rag_graph.kg_base.load_v1_knowledge_base(
+            nodes_path=node_file,
+            edges_path=edge_file,
+            template_cluster_path=tempcluster_file,
+        )
+    elif input_type == RAGVersion.V2:
+        raise ValueError("Currently only v1 input type is supported.")
+    else:
+        raise ValueError(f"Input type '{input_type}' is not supported.")
 
-# @output_on_top
-# @ops.input_position(vector_store="bottom", text_embedder="bottom")
-# @op("RAG graph")
-# def rag_graph(vector_store, text_embedder):
-#     vector_store = vector_store[0]["vector_store"]
-#     text_embedder = text_embedder[0]["text_embedder"]
-#     rag_graph = RAGGraph(
-#         PandasKnowledgeBaseGraph(vector_store=vector_store, text_embedder=text_embedder)
-#     )
-#     return {"rag_graph": rag_graph}
+    return {"rag_graph": rag_graph}
 
 
 @output_on_top
-@ops.input_position(rag_graph="bottom")
 @op("LynxScribe RAG Graph Chatbot Builder")
 @mem.cache
 def ls_rag_chatbot_builder(
-    file_urls,
     rag_graph,
     *,
     scenario_file: str = "uploads/lynx_chatbot_scenario_selector.yaml",
     node_types: str = "intent_cluster",
-    input_type: str = "v1",
 ):
     """
-    Builds up a RAG Graph-based chatbot. It could load the chatbot from
-    an existing folder (v1 or v2).
-
-    TODO: Later, we should not use these saved files, but we should build
-    up the chatbot from scratch - will be added soon). That time we will
-    add the summarizer-related parameters (LLM interface and model).
+    Builds up a RAG Graph-based chatbot (basically the loaded RAG graph +
+    a scenario selector).
 
     TODO: Later, the scenario selector can be built up synthetically from
-    the input documents - or semi-automated.
-
-    TODO: Currently, we do not affected by the embedder, as the files are
-    pre-loaded, so the text embedder should have the same model as the
-    one used in the files...
+    the input documents - or semi-automated, not just from the scenario
+    yaml.
     """
 
     scenarios = load_config(scenario_file)
     node_types = [t.strip() for t in node_types.split(",")]
 
     # handling inputs
-    file_urls = file_urls["file_urls"]
-    rag_graph = rag_graph[0]["rag_graph"]
-
-    # loading v1 knowledge base (shitty solution, but temporary)
-    if input_type == "v1":
-        node_file = [f for f in file_urls if "nodes.p" in f][0]
-        edge_file = [f for f in file_urls if "edges.p" in f][0]
-        tempcluster_file = [f for f in file_urls if "clusters.p" in f][0]
-        rag_graph.kg_base.load_v1_knowledge_base(
-            nodes_path=node_file,
-            edges_path=edge_file,
-            template_cluster_path=tempcluster_file,
-        )
-    elif input_type == "v2":
-        raise ValueError("Currently only v1 input type is supported.")
-    else:
-        raise ValueError(f"Input type '{input_type}' is not supported.")
+    # rag_graph = rag_graph[0]["rag_graph"] TODO: check why is it bad
+    rag_graph = rag_graph["rag_graph"]
 
     # loading the scenarios
     scenario_selector = ScenarioSelector(
@@ -453,18 +455,6 @@ def ls_rag_chatbot_builder(
     }
 
 
-# @output_on_top
-# @op("Scenario selector")
-# def scenario_selector(*, scenario_file: str, node_types="intent_cluster"):
-#     scenarios = load_config(scenario_file)
-#     node_types = [t.strip() for t in node_types.split(",")]
-#     scenario_selector = ScenarioSelector(
-#         scenarios=[Scenario(**scenario) for scenario in scenarios],
-#         node_types=node_types,
-#     )
-#     return {"scenario_selector": scenario_selector}
-
-
 @output_on_top
 @ops.input_position(knowledge_base="bottom", chat_processor="bottom")
 @op("LynxScribe RAG Graph Chatbot Backend")
@@ -526,35 +516,6 @@ def ls_rag_chatbot_backend(
     return {"chat_api": c}
 
 
-# @output_on_top
-# @ops.input_position(rag_graph="bottom", scenario_selector="bottom", llm="bottom")
-# @op("RAG chatbot")
-# def rag_chatbot(
-#     rag_graph,
-#     scenario_selector,
-#     llm,
-#     *,
-#     negative_answer=DEFAULT_NEGATIVE_ANSWER,
-#     limits_by_type="{}",
-#     strict_limits=True,
-#     max_results=5,
-# ):
-#     rag_graph = rag_graph[0]["rag_graph"]
-#     scenario_selector = scenario_selector[0]["scenario_selector"]
-#     llm = llm[0]["llm"]
-#     limits_by_type = json.loads(limits_by_type)
-#     rag_chatbot = RAGChatbot(
-#         rag_graph=rag_graph,
-#         scenario_selector=scenario_selector,
-#         llm=llm,
-#         negative_answer=negative_answer,
-#         limits_by_type=limits_by_type,
-#         strict_limits=strict_limits,
-#         max_results=max_results,
-#     )
-#     return {"chatbot": rag_chatbot}
-
-
 @output_on_top
 @ops.input_position(processor="bottom")
 @op("Chat processor")
@@ -622,39 +583,6 @@ def input_chat(*, chat: str):
     return {"text": chat}
 
 
-# @output_on_top
-# @ops.input_position(chatbot="bottom", chat_processor="bottom", knowledge_base="bottom")
-# @op("Chat API")
-# def chat_api(chatbot, chat_processor, knowledge_base, *, model="gpt-4o-mini"):
-#     chatbot = chatbot[0]["chatbot"]
-#     chat_processor = chat_processor[0]["chat_processor"]
-#     knowledge_base = knowledge_base[0]
-#     c = ChatAPI(
-#         chatbot=chatbot,
-#         chat_processor=chat_processor,
-#         model=model,
-#     )
-#     if knowledge_base:
-#         c.chatbot.rag_graph.kg_base.load_v1_knowledge_base(**knowledge_base)
-#         c.chatbot.scenario_selector.check_compatibility(c.chatbot.rag_graph)
-#     return {"chat_api": c}
-
-
-# @output_on_top
-# @op("Knowledge base")
-# def knowledge_base(
-#     *,
-#     nodes_path="nodes.pickle",
-#     edges_path="edges.pickle",
-#     template_cluster_path="tempclusters.pickle",
-# ):
-#     return {
-#         "nodes_path": nodes_path,
-#         "edges_path": edges_path,
-#         "template_cluster_path": template_cluster_path,
-#     }
-
-
 @op("View", view="table_view")
 def view(input):
     columns = [str(c) for c in input.keys() if not str(c).startswith("_")]
@@ -772,7 +700,7 @@ def dictionary_corrector(dict_string: str, expected_keys: list | None = None) ->
         dstring_prc = "}".join(dstring_prc.split("}")[:-1]) + "}"
 
     try:
-        trf_dict = eval(dstring_prc)
+        trf_dict = json.loads(dstring_prc)
         if expected_keys:
             for _key in expected_keys:
                 if _key in trf_dict: