Update app.py
Browse files
app.py
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
import gradio as gr
|
2 |
-
import logging, os, sys, time
|
3 |
|
4 |
from dotenv import load_dotenv, find_dotenv
|
5 |
|
@@ -7,6 +7,8 @@ from rag_langchain import LangChainRAG
|
|
7 |
from rag_llamaindex import LlamaIndexRAG
|
8 |
from trace import trace_wandb
|
9 |
|
|
|
|
|
10 |
_ = load_dotenv(find_dotenv())
|
11 |
|
12 |
RAG_INGESTION = False # load, split, embed, and store documents
|
@@ -34,55 +36,58 @@ def invoke(openai_api_key, prompt, rag_option):
|
|
34 |
if (rag_option is None):
|
35 |
raise gr.Error("Retrieval-Augmented Generation is required.")
|
36 |
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
if (
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
completion = ""
|
48 |
-
result = ""
|
49 |
-
callback = ""
|
50 |
-
err_msg = ""
|
51 |
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
rag = LangChainRAG()
|
57 |
-
completion, callback = rag.rag_chain(config, prompt)
|
58 |
-
result = completion["result"]
|
59 |
-
elif (rag_option == RAG_LLAMAINDEX):
|
60 |
-
rag = LlamaIndexRAG()
|
61 |
-
result, callback = rag.retrieval(config, prompt)
|
62 |
-
else:
|
63 |
-
rag = LangChainRAG()
|
64 |
-
completion, callback = rag.llm_chain(config, prompt)
|
65 |
-
result = completion.generations[0][0].text
|
66 |
-
except Exception as e:
|
67 |
-
err_msg = e
|
68 |
-
|
69 |
-
raise gr.Error(e)
|
70 |
-
finally:
|
71 |
-
end_time_ms = round(time.time() * 1000)
|
72 |
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
86 |
|
87 |
gr.close_all()
|
88 |
|
|
|
1 |
import gradio as gr
|
2 |
+
import logging, os, sys, threading, time
|
3 |
|
4 |
from dotenv import load_dotenv, find_dotenv
|
5 |
|
|
|
7 |
from rag_llamaindex import LlamaIndexRAG
|
8 |
from trace import trace_wandb
|
9 |
|
10 |
+
lock = threading.Lock()
|
11 |
+
|
12 |
_ = load_dotenv(find_dotenv())
|
13 |
|
14 |
RAG_INGESTION = False # load, split, embed, and store documents
|
|
|
36 |
if (rag_option is None):
|
37 |
raise gr.Error("Retrieval-Augmented Generation is required.")
|
38 |
|
39 |
+
with lock:
|
40 |
+
os.environ["OPENAI_API_KEY"] = openai_api_key
|
41 |
+
|
42 |
+
if (RAG_INGESTION):
|
43 |
+
if (rag_option == RAG_LANGCHAIN):
|
44 |
+
rag = LangChainRAG()
|
45 |
+
rag.ingestion(config)
|
46 |
+
elif (rag_option == RAG_LLAMAINDEX):
|
47 |
+
rag = LlamaIndexRAG()
|
48 |
+
rag.ingestion(config)
|
|
|
|
|
|
|
|
|
49 |
|
50 |
+
completion = ""
|
51 |
+
result = ""
|
52 |
+
callback = ""
|
53 |
+
err_msg = ""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
54 |
|
55 |
+
try:
|
56 |
+
start_time_ms = round(time.time() * 1000)
|
57 |
+
|
58 |
+
if (rag_option == RAG_LANGCHAIN):
|
59 |
+
rag = LangChainRAG()
|
60 |
+
completion, callback = rag.rag_chain(config, prompt)
|
61 |
+
result = completion["result"]
|
62 |
+
elif (rag_option == RAG_LLAMAINDEX):
|
63 |
+
rag = LlamaIndexRAG()
|
64 |
+
result, callback = rag.retrieval(config, prompt)
|
65 |
+
else:
|
66 |
+
rag = LangChainRAG()
|
67 |
+
completion, callback = rag.llm_chain(config, prompt)
|
68 |
+
result = completion.generations[0][0].text
|
69 |
+
except Exception as e:
|
70 |
+
err_msg = e
|
71 |
+
|
72 |
+
raise gr.Error(e)
|
73 |
+
finally:
|
74 |
+
end_time_ms = round(time.time() * 1000)
|
75 |
+
|
76 |
+
trace_wandb(
|
77 |
+
config,
|
78 |
+
rag_option,
|
79 |
+
prompt,
|
80 |
+
completion,
|
81 |
+
result,
|
82 |
+
callback,
|
83 |
+
err_msg,
|
84 |
+
start_time_ms,
|
85 |
+
end_time_ms
|
86 |
+
)
|
87 |
+
|
88 |
+
del os.environ["OPENAI_API_KEY"]
|
89 |
+
|
90 |
+
return result
|
91 |
|
92 |
gr.close_all()
|
93 |
|