Upload app.py
Browse files
app.py
CHANGED
@@ -34,8 +34,8 @@ client = OpenAI(
|
|
34 |
embedding_model = SentenceTransformerEmbeddings(model_name='thenlper/gte-large')
|
35 |
# Load the persisted vectorDB
|
36 |
collection_name = 'Dataset-10k'
|
37 |
-
|
38 |
-
|
39 |
collection_name=collection_name,
|
40 |
persist_directory='./dataset_db',
|
41 |
embedding_function=embedding_model
|
@@ -91,10 +91,9 @@ def predict(user_input,company):
|
|
91 |
}
|
92 |
|
93 |
filter = "dataset/"+company+"-10-k-2023.pdf"
|
94 |
-
|
95 |
-
|
96 |
# Create context_for_query
|
97 |
-
relevant_document_chunks =
|
98 |
context_list = [d.page_content for d in relevant_document_chunks]
|
99 |
context_for_query = ". ".join(context_list)
|
100 |
|
@@ -106,7 +105,7 @@ def predict(user_input,company):
|
|
106 |
)
|
107 |
}
|
108 |
]
|
109 |
-
|
110 |
# Create messages
|
111 |
try:
|
112 |
response = client.chat.completions.create(
|
@@ -117,14 +116,14 @@ def predict(user_input,company):
|
|
117 |
|
118 |
prediction = response.choices[0].message.content.strip()
|
119 |
except Exception as e:
|
120 |
-
|
121 |
prediction = f'Sorry, I encountered the following error: \n {e}'
|
122 |
-
|
123 |
|
124 |
|
125 |
# Get response from the LLM
|
126 |
prediction = response.choices[0].message.content.strip()
|
127 |
-
|
128 |
# While the prediction is made, log both the inputs and outputs to a local log file
|
129 |
# While writing to the log file, ensure that the commit scheduler is locked to avoid parallel
|
130 |
# access
|
@@ -145,9 +144,9 @@ def predict(user_input,company):
|
|
145 |
# Set-up the Gradio UI
|
146 |
user_input = gr.Textbox (label = 'Query')
|
147 |
company_input = gr.Radio(
|
148 |
-
['aws','google','IBM','Meta','msft'],
|
149 |
label = 'company'
|
150 |
-
)
|
151 |
|
152 |
model_output = gr.Textbox (label = 'Response')
|
153 |
|
@@ -162,7 +161,7 @@ model_output = gr.Textbox (label = 'Response')
|
|
162 |
demo = gr.Interface(
|
163 |
fn=predict,
|
164 |
inputs=[user_input,company_input],
|
165 |
-
outputs=
|
166 |
title="RAG on 10k-reports",
|
167 |
description="This API allows you to query on annaul reports",
|
168 |
concurrency_limit=16
|
|
|
34 |
embedding_model = SentenceTransformerEmbeddings(model_name='thenlper/gte-large')
|
35 |
# Load the persisted vectorDB
|
36 |
collection_name = 'Dataset-10k'
|
37 |
+
|
38 |
+
dataset_db = Chroma(
|
39 |
collection_name=collection_name,
|
40 |
persist_directory='./dataset_db',
|
41 |
embedding_function=embedding_model
|
|
|
91 |
}
|
92 |
|
93 |
filter = "dataset/"+company+"-10-k-2023.pdf"
|
94 |
+
|
|
|
95 |
# Create context_for_query
|
96 |
+
relevant_document_chunks = dataset_db.similarity_search(user_question, k=5, filter = {"source":"dataset/google-10-k-2023.pdf"})
|
97 |
context_list = [d.page_content for d in relevant_document_chunks]
|
98 |
context_for_query = ". ".join(context_list)
|
99 |
|
|
|
105 |
)
|
106 |
}
|
107 |
]
|
108 |
+
|
109 |
# Create messages
|
110 |
try:
|
111 |
response = client.chat.completions.create(
|
|
|
116 |
|
117 |
prediction = response.choices[0].message.content.strip()
|
118 |
except Exception as e:
|
119 |
+
|
120 |
prediction = f'Sorry, I encountered the following error: \n {e}'
|
121 |
+
|
122 |
|
123 |
|
124 |
# Get response from the LLM
|
125 |
prediction = response.choices[0].message.content.strip()
|
126 |
+
|
127 |
# While the prediction is made, log both the inputs and outputs to a local log file
|
128 |
# While writing to the log file, ensure that the commit scheduler is locked to avoid parallel
|
129 |
# access
|
|
|
144 |
# Set-up the Gradio UI
|
145 |
user_input = gr.Textbox (label = 'Query')
|
146 |
company_input = gr.Radio(
|
147 |
+
['aws','google','IBM','Meta','msft'],
|
148 |
label = 'company'
|
149 |
+
)
|
150 |
|
151 |
model_output = gr.Textbox (label = 'Response')
|
152 |
|
|
|
161 |
demo = gr.Interface(
|
162 |
fn=predict,
|
163 |
inputs=[user_input,company_input],
|
164 |
+
outputs=prediction,
|
165 |
title="RAG on 10k-reports",
|
166 |
description="This API allows you to query on annaul reports",
|
167 |
concurrency_limit=16
|