Spaces:
Sleeping
Sleeping
jocko
commited on
Commit
Β·
e146235
1
Parent(s):
fd7833c
initial commit
Browse files- requirements.txt +6 -1
- src/requirements.txt +4 -0
- src/runtime.txt +1 -0
- src/streamlit_app.py +143 -38
requirements.txt
CHANGED
@@ -1,3 +1,8 @@
|
|
1 |
altair
|
2 |
pandas
|
3 |
-
streamlit
|
|
|
|
|
|
|
|
|
|
|
|
1 |
altair
|
2 |
pandas
|
3 |
+
streamlit
|
4 |
+
torch
|
5 |
+
transformers
|
6 |
+
sentence-transformers
|
7 |
+
datasets
|
8 |
+
openai
|
src/requirements.txt
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
streamlit
|
2 |
+
openai>=1.2.0
|
3 |
+
sentence-transformers
|
4 |
+
torch
|
src/runtime.txt
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
python-3.10.12
|
src/streamlit_app.py
CHANGED
@@ -1,40 +1,145 @@
|
|
1 |
-
import
|
2 |
-
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
import streamlit as st
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
|
6 |
-
""
|
7 |
-
|
8 |
-
|
9 |
-
Edit `/streamlit_app.py` to customize this app to your heart's desire :heart:.
|
10 |
-
If you have any questions, checkout our [documentation](https://docs.streamlit.io) and [community
|
11 |
-
forums](https://discuss.streamlit.io).
|
12 |
-
|
13 |
-
In the meantime, below is an example of what you can do with just a few lines of code:
|
14 |
-
"""
|
15 |
-
|
16 |
-
num_points = st.slider("Number of points in spiral", 1, 10000, 1100)
|
17 |
-
num_turns = st.slider("Number of turns in spiral", 1, 300, 31)
|
18 |
-
|
19 |
-
indices = np.linspace(0, 1, num_points)
|
20 |
-
theta = 2 * np.pi * num_turns * indices
|
21 |
-
radius = indices
|
22 |
-
|
23 |
-
x = radius * np.cos(theta)
|
24 |
-
y = radius * np.sin(theta)
|
25 |
-
|
26 |
-
df = pd.DataFrame({
|
27 |
-
"x": x,
|
28 |
-
"y": y,
|
29 |
-
"idx": indices,
|
30 |
-
"rand": np.random.randn(num_points),
|
31 |
-
})
|
32 |
-
|
33 |
-
st.altair_chart(alt.Chart(df, height=700, width=700)
|
34 |
-
.mark_point(filled=True)
|
35 |
-
.encode(
|
36 |
-
x=alt.X("x", axis=None),
|
37 |
-
y=alt.Y("y", axis=None),
|
38 |
-
color=alt.Color("idx", legend=None, scale=alt.Scale()),
|
39 |
-
size=alt.Size("rand", legend=None, scale=alt.Scale(range=[1, 150])),
|
40 |
-
))
|
|
|
1 |
+
import os
|
2 |
+
|
3 |
+
# β
Set all relevant cache directories to a writable location
|
4 |
+
os.environ["HF_HOME"] = "/tmp/cache"
|
5 |
+
os.environ["TRANSFORMERS_CACHE"] = "/tmp/cache/transformers"
|
6 |
+
os.environ["SENTENCE_TRANSFORMERS_HOME"] = "/tmp/cache/sentence_transformers"
|
7 |
+
os.environ["HF_DATASETS_CACHE"] = "/tmp/cache/hf_datasets"
|
8 |
+
os.environ["TORCH_HOME"] = "/tmp/cache/torch"
|
9 |
+
|
10 |
+
# β
Create the directories if they don't exist
|
11 |
+
for path in [
|
12 |
+
"/tmp/cache",
|
13 |
+
"/tmp/cache/transformers",
|
14 |
+
"/tmp/cache/sentence_transformers",
|
15 |
+
"/tmp/cache/hf_datasets",
|
16 |
+
"/tmp/cache/torch"
|
17 |
+
]:
|
18 |
+
os.makedirs(path, exist_ok=True)
|
19 |
+
import json
|
20 |
+
import torch
|
21 |
+
import openai
|
22 |
+
import os
|
23 |
+
from sentence_transformers import SentenceTransformer, util
|
24 |
import streamlit as st
|
25 |
+
from pathlib import Path
|
26 |
+
|
27 |
+
# === CONFIG ===
|
28 |
+
# Set the API key
|
29 |
+
client = openai.OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
|
30 |
+
#openai.api_key = os.getenv("OPENAI_API_KEY")
|
31 |
+
# REMEDI_PATH = "ReMeDi-base.json"
|
32 |
+
BASE_DIR = Path(__file__).parent
|
33 |
+
REMEDI_PATH = BASE_DIR / "ReMeDi-base.json"
|
34 |
+
|
35 |
+
# Check if file exists
|
36 |
+
if not REMEDI_PATH.exists():
|
37 |
+
raise FileNotFoundError(f"β File not found: {REMEDI_PATH}")
|
38 |
+
|
39 |
+
# Load the file
|
40 |
+
with open(REMEDI_PATH, "r", encoding="utf-8") as f:
|
41 |
+
data = json.load(f)
|
42 |
+
|
43 |
+
# === LOAD MODEL ===
|
44 |
+
@st.cache_resource
|
45 |
+
def load_model():
|
46 |
+
return SentenceTransformer("all-MiniLM-L6-v2")
|
47 |
+
#return model
|
48 |
+
|
49 |
+
@st.cache_resource
|
50 |
+
def load_data():
|
51 |
+
with open(REMEDI_PATH, "r", encoding="utf-8") as f:
|
52 |
+
data = json.load(f)
|
53 |
+
dialogue_pairs = []
|
54 |
+
for conversation in data:
|
55 |
+
turns = conversation["information"]
|
56 |
+
for i in range(len(turns)-1):
|
57 |
+
if turns[i]["role"] == "patient" and turns[i+1]["role"] == "doctor":
|
58 |
+
dialogue_pairs.append({
|
59 |
+
"patient": turns[i]["sentence"],
|
60 |
+
"doctor": turns[i+1]["sentence"]
|
61 |
+
})
|
62 |
+
return dialogue_pairs
|
63 |
+
|
64 |
+
@st.cache_data
|
65 |
+
def build_embeddings(dialogue_pairs, _model):
|
66 |
+
patient_sentences = [pair["patient"] for pair in dialogue_pairs]
|
67 |
+
embeddings = _model.encode(patient_sentences, convert_to_tensor=True)
|
68 |
+
return embeddings
|
69 |
+
|
70 |
+
# === TRANSLATE USING GPT ===
|
71 |
+
def translate_to_english(chinese_text):
|
72 |
+
prompt = f"Translate the following Chinese medical response to English:\n\n{chinese_text}"
|
73 |
+
try:
|
74 |
+
response = client.chat.completions.create(
|
75 |
+
model="gpt-4",
|
76 |
+
messages=[{"role": "user", "content": prompt}],
|
77 |
+
temperature=0.2
|
78 |
+
)
|
79 |
+
return response.choices[0].message.content
|
80 |
+
|
81 |
+
#return response.choices[0].message["content"].strip()
|
82 |
+
except Exception as e:
|
83 |
+
return f"Translation failed: {str(e)}"
|
84 |
+
|
85 |
+
def gpt_direct_response(user_input):
|
86 |
+
prompt = f"You are a knowledgeable and compassionate medical assistant. Answer the following patient question clearly and concisely:\n\n{user_input}"
|
87 |
+
try:
|
88 |
+
response = client.chat.completions.create(
|
89 |
+
model="gpt-4", # or "gpt-3.5-turbo" to save credits
|
90 |
+
messages=[{"role": "user", "content": prompt}],
|
91 |
+
temperature=0.5
|
92 |
+
)
|
93 |
+
return response.choices[0].message.content
|
94 |
+
except Exception as e:
|
95 |
+
return f"GPT response failed: {str(e)}"
|
96 |
+
|
97 |
+
|
98 |
+
# === CHATBOT FUNCTION ===
|
99 |
+
def chatbot_response(user_input, _model, dialogue_pairs, patient_embeddings, top_k=1):
|
100 |
+
user_embedding = _model.encode(user_input, convert_to_tensor=True)
|
101 |
+
similarities = util.cos_sim(user_embedding, patient_embeddings)[0]
|
102 |
+
top_idx = torch.topk(similarities, k=top_k).indices[0].item()
|
103 |
+
|
104 |
+
match = dialogue_pairs[top_idx]
|
105 |
+
translated = translate_to_english(match["doctor"])
|
106 |
+
|
107 |
+
return {
|
108 |
+
"matched_question": match["patient"],
|
109 |
+
"original_response": match["doctor"],
|
110 |
+
"translated_response": translated
|
111 |
+
}
|
112 |
+
|
113 |
+
# === MAIN APP ===
|
114 |
+
st.set_page_config(page_title="Dr_Q_bot", layout="centered")
|
115 |
+
st.title("π©Ί Dr_Q_bot - Medical Chatbot")
|
116 |
+
st.write("Ask about a symptom and get an example doctor response (translated from Chinese).")
|
117 |
+
|
118 |
+
# Load resources
|
119 |
+
model = load_model()
|
120 |
+
dialogue_pairs = load_data()
|
121 |
+
patient_embeddings = build_embeddings(dialogue_pairs, model)
|
122 |
+
|
123 |
+
# Chat UI
|
124 |
+
user_input = st.text_input("Describe your symptom:")
|
125 |
+
|
126 |
+
if st.button("Submit") and user_input:
|
127 |
+
with st.spinner("Thinking..."):
|
128 |
+
result = chatbot_response(user_input, model, dialogue_pairs, patient_embeddings)
|
129 |
+
gpt_response = gpt_direct_response(user_input)
|
130 |
+
|
131 |
+
st.markdown("### π§ββοΈ Closest Patient Question")
|
132 |
+
st.write(result["matched_question"])
|
133 |
+
|
134 |
+
st.markdown("### π¨π³ Original Doctor Response (Chinese)")
|
135 |
+
st.write(result["original_response"])
|
136 |
+
|
137 |
+
st.markdown("### π Translated Doctor Response (English)")
|
138 |
+
st.success(result["translated_response"])
|
139 |
+
|
140 |
+
st.markdown("### π¬ GPT Doctor Response (AI-generated)")
|
141 |
+
st.info(gpt_response)
|
142 |
+
|
143 |
|
144 |
+
st.markdown("---")
|
145 |
+
st.warning("This chatbot uses real dialogue data for research and educational use only. Not a substitute for professional medical advice.")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|