Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -4,6 +4,10 @@ from openai import OpenAI
|
|
4 |
import os
|
5 |
import numpy as np
|
6 |
from sklearn.metrics.pairwise import cosine_similarity
|
|
|
|
|
|
|
|
|
7 |
|
8 |
# Load the NASA-specific bi-encoder model and tokenizer
|
9 |
bi_encoder_model_name = "nasa-impact/nasa-smd-ibm-st-v2"
|
@@ -14,27 +18,55 @@ bi_model = AutoModel.from_pretrained(bi_encoder_model_name)
|
|
14 |
api_key = os.getenv('OPENAI_API_KEY')
|
15 |
client = OpenAI(api_key=api_key)
|
16 |
|
|
|
|
|
|
|
17 |
# Define a system message to introduce Exos
|
18 |
system_message = "You are Exos, a helpful assistant specializing in Exoplanet research. Provide detailed and accurate responses related to Exoplanet research."
|
19 |
|
20 |
def encode_text(text):
|
21 |
inputs = bi_tokenizer(text, return_tensors='pt', padding=True, truncation=True, max_length=128)
|
22 |
outputs = bi_model(**inputs)
|
23 |
-
return outputs.last_hidden_state.mean(dim=1).detach().numpy().flatten()
|
24 |
|
25 |
def retrieve_relevant_context(user_input, context_texts):
|
26 |
user_embedding = encode_text(user_input).reshape(1, -1)
|
27 |
context_embeddings = np.array([encode_text(text) for text in context_texts])
|
28 |
-
context_embeddings = context_embeddings.reshape(len(context_embeddings), -1)
|
29 |
similarities = cosine_similarity(user_embedding, context_embeddings).flatten()
|
30 |
most_relevant_idx = np.argmax(similarities)
|
31 |
return context_texts[most_relevant_idx]
|
32 |
|
33 |
-
def
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
34 |
if relevant_context:
|
35 |
-
combined_input = f"Context: {relevant_context}\nQuestion: {user_input}\nAnswer:"
|
36 |
else:
|
37 |
-
combined_input = f"Question: {user_input}\nAnswer:"
|
38 |
|
39 |
response = client.chat.completions.create(
|
40 |
model="gpt-4-turbo",
|
@@ -48,23 +80,84 @@ def generate_response(user_input, relevant_context="", max_tokens=150, temperatu
|
|
48 |
frequency_penalty=frequency_penalty,
|
49 |
presence_penalty=presence_penalty
|
50 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
51 |
return response.choices[0].message.content.strip()
|
52 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
53 |
def chatbot(user_input, context="", use_encoder=False, max_tokens=150, temperature=0.7, top_p=0.9, frequency_penalty=0.5, presence_penalty=0.0):
|
54 |
if use_encoder and context:
|
55 |
context_texts = context.split("\n")
|
56 |
relevant_context = retrieve_relevant_context(user_input, context_texts)
|
57 |
else:
|
58 |
relevant_context = ""
|
59 |
-
response = generate_response(user_input, relevant_context, max_tokens, temperature, top_p, frequency_penalty, presence_penalty)
|
60 |
-
return response
|
61 |
|
62 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
63 |
iface = gr.Interface(
|
64 |
fn=chatbot,
|
65 |
inputs=[
|
66 |
-
gr.Textbox(lines=2, placeholder="
|
67 |
-
gr.Textbox(lines=5, placeholder="Enter context here
|
68 |
gr.Checkbox(label="Use NASA SMD Bi-Encoder for Context"),
|
69 |
gr.Slider(50, 1000, value=150, step=10, label="Max Tokens"),
|
70 |
gr.Slider(0.0, 1.0, value=0.7, step=0.1, label="Temperature"),
|
@@ -72,10 +165,14 @@ iface = gr.Interface(
|
|
72 |
gr.Slider(0.0, 1.0, value=0.5, step=0.1, label="Frequency Penalty"),
|
73 |
gr.Slider(0.0, 1.0, value=0.0, step=0.1, label="Presence Penalty")
|
74 |
],
|
75 |
-
outputs=
|
76 |
-
|
77 |
-
|
|
|
|
|
|
|
|
|
|
|
78 |
)
|
79 |
|
80 |
-
# Launch the interface
|
81 |
iface.launch(share=True)
|
|
|
4 |
import os
|
5 |
import numpy as np
|
6 |
from sklearn.metrics.pairwise import cosine_similarity
|
7 |
+
from docx import Document
|
8 |
+
import io
|
9 |
+
import tempfile
|
10 |
+
from astroquery.nasa_ads import ADS
|
11 |
|
12 |
# Load the NASA-specific bi-encoder model and tokenizer
|
13 |
bi_encoder_model_name = "nasa-impact/nasa-smd-ibm-st-v2"
|
|
|
18 |
api_key = os.getenv('OPENAI_API_KEY')
|
19 |
client = OpenAI(api_key=api_key)
|
20 |
|
21 |
+
# Set up NASA ADS token
|
22 |
+
ADS.TOKEN = os.getenv('ADS_API_KEY') # Ensure your ADS API key is stored in environment variables
|
23 |
+
|
24 |
# Define a system message to introduce Exos
|
25 |
system_message = "You are Exos, a helpful assistant specializing in Exoplanet research. Provide detailed and accurate responses related to Exoplanet research."
|
26 |
|
27 |
def encode_text(text):
|
28 |
inputs = bi_tokenizer(text, return_tensors='pt', padding=True, truncation=True, max_length=128)
|
29 |
outputs = bi_model(**inputs)
|
30 |
+
return outputs.last_hidden_state.mean(dim=1).detach().numpy().flatten()
|
31 |
|
32 |
def retrieve_relevant_context(user_input, context_texts):
|
33 |
user_embedding = encode_text(user_input).reshape(1, -1)
|
34 |
context_embeddings = np.array([encode_text(text) for text in context_texts])
|
35 |
+
context_embeddings = context_embeddings.reshape(len(context_embeddings), -1)
|
36 |
similarities = cosine_similarity(user_embedding, context_embeddings).flatten()
|
37 |
most_relevant_idx = np.argmax(similarities)
|
38 |
return context_texts[most_relevant_idx]
|
39 |
|
40 |
+
def fetch_nasa_ads_references(prompt):
|
41 |
+
try:
|
42 |
+
# Use the entire prompt for the query
|
43 |
+
simplified_query = prompt
|
44 |
+
|
45 |
+
# Query NASA ADS for relevant papers
|
46 |
+
papers = ADS.query_simple(simplified_query)
|
47 |
+
|
48 |
+
if not papers or len(papers) == 0:
|
49 |
+
return [("No results found", "N/A", "N/A")]
|
50 |
+
|
51 |
+
# Include authors in the references
|
52 |
+
references = [
|
53 |
+
(
|
54 |
+
paper['title'][0],
|
55 |
+
", ".join(paper['author'][:3]) + (" et al." if len(paper['author']) > 3 else ""),
|
56 |
+
paper['bibcode']
|
57 |
+
)
|
58 |
+
for paper in papers[:5] # Limit to 5 references
|
59 |
+
]
|
60 |
+
return references
|
61 |
+
|
62 |
+
except Exception as e:
|
63 |
+
return [("Error fetching references", str(e), "N/A")]
|
64 |
+
|
65 |
+
def generate_response(user_input, relevant_context="", references=[], max_tokens=150, temperature=0.7, top_p=0.9, frequency_penalty=0.5, presence_penalty=0.0):
|
66 |
if relevant_context:
|
67 |
+
combined_input = f"Context: {relevant_context}\nQuestion: {user_input}\nAnswer (please organize the answer in a structured format with topics and subtopics):"
|
68 |
else:
|
69 |
+
combined_input = f"Question: {user_input}\nAnswer (please organize the answer in a structured format with topics and subtopics):"
|
70 |
|
71 |
response = client.chat.completions.create(
|
72 |
model="gpt-4-turbo",
|
|
|
80 |
frequency_penalty=frequency_penalty,
|
81 |
presence_penalty=presence_penalty
|
82 |
)
|
83 |
+
|
84 |
+
# Append references to the response
|
85 |
+
if references:
|
86 |
+
response_content = response.choices[0].message.content.strip()
|
87 |
+
references_text = "\n\nADS References:\n" + "\n".join(
|
88 |
+
[f"- {title} by {authors} (Bibcode: {bibcode})" for title, authors, bibcode in references]
|
89 |
+
)
|
90 |
+
return f"{response_content}\n{references_text}"
|
91 |
+
|
92 |
return response.choices[0].message.content.strip()
|
93 |
|
94 |
+
def export_to_word(response_content):
|
95 |
+
doc = Document()
|
96 |
+
doc.add_heading('AI Generated SCDD', 0)
|
97 |
+
for line in response_content.split('\n'):
|
98 |
+
doc.add_paragraph(line)
|
99 |
+
|
100 |
+
temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".docx")
|
101 |
+
doc.save(temp_file.name)
|
102 |
+
|
103 |
+
return temp_file.name
|
104 |
+
|
105 |
def chatbot(user_input, context="", use_encoder=False, max_tokens=150, temperature=0.7, top_p=0.9, frequency_penalty=0.5, presence_penalty=0.0):
|
106 |
if use_encoder and context:
|
107 |
context_texts = context.split("\n")
|
108 |
relevant_context = retrieve_relevant_context(user_input, context_texts)
|
109 |
else:
|
110 |
relevant_context = ""
|
|
|
|
|
111 |
|
112 |
+
# Fetch NASA ADS references using the full prompt
|
113 |
+
references = fetch_nasa_ads_references(user_input)
|
114 |
+
|
115 |
+
# Generate response from GPT-4
|
116 |
+
response = generate_response(user_input, relevant_context, references, max_tokens, temperature, top_p, frequency_penalty, presence_penalty)
|
117 |
+
|
118 |
+
# Export the response to a Word document
|
119 |
+
word_doc_path = export_to_word(response)
|
120 |
+
|
121 |
+
# Embed Miro iframe
|
122 |
+
iframe_html = """
|
123 |
+
<iframe width="768" height="432" src="https://miro.com/app/live-embed/uXjVKuVTcF8=/?moveToViewport=-331,-462,5434,3063&embedId=710273023721" frameborder="0" scrolling="no" allow="fullscreen; clipboard-read; clipboard-write" allowfullscreen></iframe>
|
124 |
+
"""
|
125 |
+
|
126 |
+
mapify_button_html = """
|
127 |
+
<style>
|
128 |
+
.mapify-button {
|
129 |
+
background: linear-gradient(135deg, #1E90FF 0%, #87CEFA 100%);
|
130 |
+
border: none;
|
131 |
+
color: white;
|
132 |
+
padding: 15px 35px;
|
133 |
+
text-align: center;
|
134 |
+
text-decoration: none;
|
135 |
+
display: inline-block;
|
136 |
+
font-size: 18px;
|
137 |
+
font-weight: bold;
|
138 |
+
margin: 20px 2px;
|
139 |
+
cursor: pointer;
|
140 |
+
border-radius: 25px;
|
141 |
+
transition: all 0.3s ease;
|
142 |
+
box-shadow: 0 4px 15px rgba(0, 0, 0, 0.2);
|
143 |
+
}
|
144 |
+
.mapify-button:hover {
|
145 |
+
background: linear-gradient(135deg, #4682B4 0%, #1E90FF 100%);
|
146 |
+
box-shadow: 0 6px 20px rgba(0, 0, 0, 0.3);
|
147 |
+
transform: scale(1.05);
|
148 |
+
}
|
149 |
+
</style>
|
150 |
+
<a href="https://mapify.so/app/new" target="_blank">
|
151 |
+
<button class="mapify-button">Create Mind Map on Mapify</button>
|
152 |
+
</a>
|
153 |
+
"""
|
154 |
+
return response, iframe_html, mapify_button_html, word_doc_path
|
155 |
+
|
156 |
iface = gr.Interface(
|
157 |
fn=chatbot,
|
158 |
inputs=[
|
159 |
+
gr.Textbox(lines=2, placeholder="Formulate your science goal...", label="Prompt"),
|
160 |
+
gr.Textbox(lines=5, placeholder="Enter some context here...", label="Context"),
|
161 |
gr.Checkbox(label="Use NASA SMD Bi-Encoder for Context"),
|
162 |
gr.Slider(50, 1000, value=150, step=10, label="Max Tokens"),
|
163 |
gr.Slider(0.0, 1.0, value=0.7, step=0.1, label="Temperature"),
|
|
|
165 |
gr.Slider(0.0, 1.0, value=0.5, step=0.1, label="Frequency Penalty"),
|
166 |
gr.Slider(0.0, 1.0, value=0.0, step=0.1, label="Presence Penalty")
|
167 |
],
|
168 |
+
outputs=[
|
169 |
+
gr.Textbox(label="Model Response..."),
|
170 |
+
gr.HTML(label="Miro"),
|
171 |
+
gr.HTML(label="Generate Mind Map on Mapify"),
|
172 |
+
gr.File(label="Download SCDD", type="filepath"),
|
173 |
+
],
|
174 |
+
title="SCDDBot - NASA SMD SCDD AI Assistant [version-0.2a]",
|
175 |
+
description="SCDDBot is an AI-powered assistant for generating and visualising HWO Science Cases",
|
176 |
)
|
177 |
|
|
|
178 |
iface.launch(share=True)
|