Spaces:
Build error
Build error
Update app.py
Browse files
app.py
CHANGED
|
@@ -18,7 +18,7 @@ llm = HuggingFaceHub(repo_id="mistralai/Mistral-7B-Instruct-v0.3", model_kwargs=
|
|
| 18 |
st.set_page_config(page_title="DocuMentorAI", layout="wide", page_icon="📄")
|
| 19 |
st.title("📄 DocuMentorAI")
|
| 20 |
|
| 21 |
-
# Improved CSS
|
| 22 |
st.markdown("""
|
| 23 |
<style>
|
| 24 |
.output-container {
|
|
@@ -26,14 +26,13 @@ st.markdown("""
|
|
| 26 |
padding: 20px;
|
| 27 |
border-radius: 10px;
|
| 28 |
margin-top: 20px;
|
| 29 |
-
}
|
| 30 |
-
.generated-content {
|
| 31 |
-
font-size: 16px;
|
| 32 |
-
line-height: 1.6;
|
| 33 |
white-space: pre-wrap;
|
| 34 |
}
|
| 35 |
-
.
|
| 36 |
-
|
|
|
|
|
|
|
|
|
|
| 37 |
}
|
| 38 |
</style>
|
| 39 |
""", unsafe_allow_html=True)
|
|
@@ -63,7 +62,7 @@ def extract_text(uploaded_file):
|
|
| 63 |
|
| 64 |
def extract_professor_details(text):
|
| 65 |
professor_pattern = r"(Dr\.|Professor|Prof\.?)\s+([A-Z][a-z]+\s[A-Z][a-z]+)"
|
| 66 |
-
university_pattern = r"(University|Institute|College|School of [A-Za-z]+)"
|
| 67 |
|
| 68 |
professor_match = re.search(professor_pattern, text)
|
| 69 |
university_match = re.search(university_pattern, text)
|
|
@@ -71,14 +70,87 @@ def extract_professor_details(text):
|
|
| 71 |
return (professor_match.group(0) if professor_match else "Not Found",
|
| 72 |
university_match.group(0) if university_match else "Not Found")
|
| 73 |
|
| 74 |
-
|
| 75 |
-
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
|
| 81 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 82 |
if 'generated_content' not in st.session_state:
|
| 83 |
st.session_state.generated_content = {
|
| 84 |
'email': None,
|
|
@@ -90,43 +162,68 @@ if 'generated_content' not in st.session_state:
|
|
| 90 |
# Template Definitions
|
| 91 |
templates = {
|
| 92 |
'email': PromptTemplate.from_template("""
|
| 93 |
-
Write a
|
| 94 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 95 |
Professor: {professor_name}
|
| 96 |
University: {university_name}
|
| 97 |
Research Interests: {research_interests}
|
| 98 |
Why This Lab: {reason}
|
| 99 |
-
CV Highlights: {resume_text}
|
| 100 |
"""),
|
|
|
|
| 101 |
'cover_letter': PromptTemplate.from_template("""
|
| 102 |
-
Write a
|
| 103 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 104 |
Job Title: {job_title}
|
| 105 |
Company: {company}
|
| 106 |
Key Skills: {key_skills}
|
| 107 |
-
|
| 108 |
"""),
|
|
|
|
| 109 |
'research_statement': PromptTemplate.from_template("""
|
| 110 |
-
Write a research statement
|
| 111 |
-
|
| 112 |
-
|
| 113 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 114 |
Future Goals: {future_goals}
|
| 115 |
"""),
|
|
|
|
| 116 |
'sop': PromptTemplate.from_template("""
|
| 117 |
-
Write a
|
| 118 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 119 |
Motivation: {motivation}
|
| 120 |
Academic Background: {academic_background}
|
| 121 |
-
Research
|
| 122 |
Career Goals: {career_goals}
|
| 123 |
-
|
| 124 |
""")
|
| 125 |
}
|
| 126 |
|
| 127 |
# Create LangChain instances
|
| 128 |
chains = {key: LLMChain(llm=llm, prompt=template) for key, template in templates.items()}
|
| 129 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 130 |
# Tab Layout
|
| 131 |
tab1, tab2, tab3, tab4 = st.tabs(["Cold Email", "Cover Letter", "Research Statement", "SOP"])
|
| 132 |
|
|
@@ -140,13 +237,14 @@ with tab1:
|
|
| 140 |
if job_opening_text and cv_resume_text:
|
| 141 |
with st.spinner("Generating..."):
|
| 142 |
try:
|
| 143 |
-
|
| 144 |
"professor_name": professor_name,
|
| 145 |
"university_name": university_name,
|
| 146 |
"research_interests": research_interests,
|
| 147 |
"reason": reason,
|
| 148 |
"resume_text": cv_resume_text
|
| 149 |
})
|
|
|
|
| 150 |
except Exception as e:
|
| 151 |
st.error(f"Generation error: {e}")
|
| 152 |
else:
|
|
@@ -155,7 +253,7 @@ with tab1:
|
|
| 155 |
if st.session_state.generated_content['email']:
|
| 156 |
st.markdown('<div class="output-container">', unsafe_allow_html=True)
|
| 157 |
st.markdown(st.session_state.generated_content['email'])
|
| 158 |
-
st.download_button("Download Email", st.session_state.generated_content['email'],
|
| 159 |
file_name="cold_email.txt", key="email_download")
|
| 160 |
st.markdown('</div>', unsafe_allow_html=True)
|
| 161 |
|
|
@@ -169,12 +267,13 @@ with tab2:
|
|
| 169 |
if job_opening_text and cv_resume_text:
|
| 170 |
with st.spinner("Generating..."):
|
| 171 |
try:
|
| 172 |
-
|
| 173 |
"job_title": job_title,
|
| 174 |
"company": company_name,
|
| 175 |
"key_skills": key_skills,
|
| 176 |
"resume_text": cv_resume_text
|
| 177 |
})
|
|
|
|
| 178 |
except Exception as e:
|
| 179 |
st.error(f"Generation error: {e}")
|
| 180 |
else:
|
|
@@ -196,11 +295,12 @@ with tab3:
|
|
| 196 |
if st.button("Generate Research Statement", key="research_stmt_btn"):
|
| 197 |
with st.spinner("Generating..."):
|
| 198 |
try:
|
| 199 |
-
|
| 200 |
"research_background": research_background,
|
| 201 |
"key_projects": key_projects,
|
| 202 |
"future_goals": future_goals
|
| 203 |
})
|
|
|
|
| 204 |
except Exception as e:
|
| 205 |
st.error(f"Generation error: {e}")
|
| 206 |
|
|
@@ -222,13 +322,14 @@ with tab4:
|
|
| 222 |
if st.button("Generate SOP", key="sop_btn"):
|
| 223 |
with st.spinner("Generating..."):
|
| 224 |
try:
|
| 225 |
-
|
| 226 |
"motivation": motivation,
|
| 227 |
"academic_background": academic_background,
|
| 228 |
"research_experiences": research_experiences,
|
| 229 |
"career_goals": career_goals,
|
| 230 |
"why_this_program": why_this_program
|
| 231 |
})
|
|
|
|
| 232 |
except Exception as e:
|
| 233 |
st.error(f"Generation error: {e}")
|
| 234 |
|
|
|
|
| 18 |
st.set_page_config(page_title="DocuMentorAI", layout="wide", page_icon="📄")
|
| 19 |
st.title("📄 DocuMentorAI")
|
| 20 |
|
| 21 |
+
# Improved CSS
|
| 22 |
st.markdown("""
|
| 23 |
<style>
|
| 24 |
.output-container {
|
|
|
|
| 26 |
padding: 20px;
|
| 27 |
border-radius: 10px;
|
| 28 |
margin-top: 20px;
|
|
|
|
|
|
|
|
|
|
|
|
|
| 29 |
white-space: pre-wrap;
|
| 30 |
}
|
| 31 |
+
.stTextArea textarea {
|
| 32 |
+
font-size: 16px !important;
|
| 33 |
+
}
|
| 34 |
+
.stButton button {
|
| 35 |
+
width: 100%;
|
| 36 |
}
|
| 37 |
</style>
|
| 38 |
""", unsafe_allow_html=True)
|
|
|
|
| 62 |
|
| 63 |
def extract_professor_details(text):
|
| 64 |
professor_pattern = r"(Dr\.|Professor|Prof\.?)\s+([A-Z][a-z]+\s[A-Z][a-z]+)"
|
| 65 |
+
university_pattern = r"(University|Institute|College|School of [A-Z][A-Za-z\s]+)"
|
| 66 |
|
| 67 |
professor_match = re.search(professor_pattern, text)
|
| 68 |
university_match = re.search(university_pattern, text)
|
|
|
|
| 70 |
return (professor_match.group(0) if professor_match else "Not Found",
|
| 71 |
university_match.group(0) if university_match else "Not Found")
|
| 72 |
|
| 73 |
+
def clean_email_output(email_text):
|
| 74 |
+
"""Clean and format email content"""
|
| 75 |
+
start_idx = email_text.find("Dear")
|
| 76 |
+
if start_idx == -1:
|
| 77 |
+
start_idx = 0
|
| 78 |
+
|
| 79 |
+
end_markers = ["Best regards,", "Sincerely,", "Yours sincerely,", "Kind regards,"]
|
| 80 |
+
end_idx = len(email_text)
|
| 81 |
+
for marker in end_markers:
|
| 82 |
+
idx = email_text.find(marker)
|
| 83 |
+
if idx != -1:
|
| 84 |
+
end_idx = email_text.find("\n\n", idx) if email_text.find("\n\n", idx) != -1 else len(email_text)
|
| 85 |
+
break
|
| 86 |
+
|
| 87 |
+
email_content = email_text[start_idx:end_idx].strip()
|
| 88 |
+
|
| 89 |
+
if "Phone:" in email_text or "Email:" in email_text:
|
| 90 |
+
contact_info = "\n\n" + "\n".join([
|
| 91 |
+
line for line in email_text[end_idx:].split("\n")
|
| 92 |
+
if any(info in line for info in ["Phone:", "Email:"])
|
| 93 |
+
]).strip()
|
| 94 |
+
email_content += contact_info
|
| 95 |
+
|
| 96 |
+
return email_content
|
| 97 |
+
|
| 98 |
+
def clean_cover_letter_output(letter_text):
|
| 99 |
+
"""Clean and format cover letter content"""
|
| 100 |
+
start_markers = ["Dear", "To Whom", "Hiring"]
|
| 101 |
+
start_idx = len(letter_text)
|
| 102 |
+
for marker in start_markers:
|
| 103 |
+
idx = letter_text.find(marker)
|
| 104 |
+
if idx != -1:
|
| 105 |
+
start_idx = min(start_idx, idx)
|
| 106 |
+
|
| 107 |
+
end_markers = ["Sincerely,", "Best regards,", "Yours truly,", "Regards,"]
|
| 108 |
+
end_idx = len(letter_text)
|
| 109 |
+
for marker in end_markers:
|
| 110 |
+
idx = letter_text.find(marker)
|
| 111 |
+
if idx != -1:
|
| 112 |
+
end_idx = letter_text.find("\n\n", idx) if letter_text.find("\n\n", idx) != -1 else len(letter_text)
|
| 113 |
+
break
|
| 114 |
+
|
| 115 |
+
return letter_text[start_idx:end_idx].strip()
|
| 116 |
+
|
| 117 |
+
def clean_research_statement_output(statement_text):
|
| 118 |
+
"""Clean and format research statement content"""
|
| 119 |
+
# Remove common headers
|
| 120 |
+
headers = ["Research Statement", "Statement of Research", "Research Interests"]
|
| 121 |
+
cleaned_text = statement_text
|
| 122 |
+
for header in headers:
|
| 123 |
+
if cleaned_text.startswith(header):
|
| 124 |
+
cleaned_text = cleaned_text[len(header):].lstrip(":\n")
|
| 125 |
+
|
| 126 |
+
# Remove any trailing references or bibliography sections
|
| 127 |
+
end_markers = ["References", "Bibliography", "Citations"]
|
| 128 |
+
for marker in end_markers:
|
| 129 |
+
idx = cleaned_text.find(marker)
|
| 130 |
+
if idx != -1:
|
| 131 |
+
cleaned_text = cleaned_text[:idx].strip()
|
| 132 |
+
|
| 133 |
+
return cleaned_text.strip()
|
| 134 |
|
| 135 |
+
def clean_sop_output(sop_text):
|
| 136 |
+
"""Clean and format Statement of Purpose content"""
|
| 137 |
+
# Remove common headers
|
| 138 |
+
headers = ["Statement of Purpose", "Personal Statement", "Academic Statement"]
|
| 139 |
+
cleaned_text = sop_text
|
| 140 |
+
for header in headers:
|
| 141 |
+
if cleaned_text.startswith(header):
|
| 142 |
+
cleaned_text = cleaned_text[len(header):].lstrip(":\n")
|
| 143 |
+
|
| 144 |
+
# Remove any trailing sections
|
| 145 |
+
end_markers = ["Thank you", "References", "Additional Information"]
|
| 146 |
+
for marker in end_markers:
|
| 147 |
+
idx = cleaned_text.find(marker)
|
| 148 |
+
if idx != -1:
|
| 149 |
+
cleaned_text = cleaned_text[:idx].strip()
|
| 150 |
+
|
| 151 |
+
return cleaned_text.strip()
|
| 152 |
+
|
| 153 |
+
# Initialize session state
|
| 154 |
if 'generated_content' not in st.session_state:
|
| 155 |
st.session_state.generated_content = {
|
| 156 |
'email': None,
|
|
|
|
| 162 |
# Template Definitions
|
| 163 |
templates = {
|
| 164 |
'email': PromptTemplate.from_template("""
|
| 165 |
+
Write ONLY a formal cold email for a research position.
|
| 166 |
+
Start with 'Dear Professor' and end with a signature.
|
| 167 |
+
Include only 2-3 most relevant highlights from the CV.
|
| 168 |
+
Do not include any other text or formatting.
|
| 169 |
+
|
| 170 |
+
Details to use:
|
| 171 |
Professor: {professor_name}
|
| 172 |
University: {university_name}
|
| 173 |
Research Interests: {research_interests}
|
| 174 |
Why This Lab: {reason}
|
|
|
|
| 175 |
"""),
|
| 176 |
+
|
| 177 |
'cover_letter': PromptTemplate.from_template("""
|
| 178 |
+
Write ONLY a professional cover letter.
|
| 179 |
+
Start with a formal greeting and end with a signature.
|
| 180 |
+
Focus on relevant experience and skills.
|
| 181 |
+
Do not include any headers or additional formatting.
|
| 182 |
+
|
| 183 |
+
Details to use:
|
| 184 |
Job Title: {job_title}
|
| 185 |
Company: {company}
|
| 186 |
Key Skills: {key_skills}
|
| 187 |
+
Relevant Experience: {resume_text}
|
| 188 |
"""),
|
| 189 |
+
|
| 190 |
'research_statement': PromptTemplate.from_template("""
|
| 191 |
+
Write ONLY a research statement.
|
| 192 |
+
Focus on research background, achievements, and future goals.
|
| 193 |
+
Maintain a professional and academic tone.
|
| 194 |
+
Do not include any headers or section titles.
|
| 195 |
+
|
| 196 |
+
Content to include:
|
| 197 |
+
Background: {research_background}
|
| 198 |
+
Key Projects: {key_projects}
|
| 199 |
Future Goals: {future_goals}
|
| 200 |
"""),
|
| 201 |
+
|
| 202 |
'sop': PromptTemplate.from_template("""
|
| 203 |
+
Write ONLY a Statement of Purpose.
|
| 204 |
+
Focus on academic journey, research interests, and career goals.
|
| 205 |
+
Maintain a personal yet professional tone.
|
| 206 |
+
Do not include any headers or section titles.
|
| 207 |
+
|
| 208 |
+
Details to use:
|
| 209 |
Motivation: {motivation}
|
| 210 |
Academic Background: {academic_background}
|
| 211 |
+
Research Experience: {research_experiences}
|
| 212 |
Career Goals: {career_goals}
|
| 213 |
+
Program Fit: {why_this_program}
|
| 214 |
""")
|
| 215 |
}
|
| 216 |
|
| 217 |
# Create LangChain instances
|
| 218 |
chains = {key: LLMChain(llm=llm, prompt=template) for key, template in templates.items()}
|
| 219 |
|
| 220 |
+
# Sidebar for Input Collection
|
| 221 |
+
with st.sidebar:
|
| 222 |
+
st.subheader("📝 Input Details")
|
| 223 |
+
job_opening_text = st.text_area("Job/Research Opening Details", height=150)
|
| 224 |
+
cv_resume_file = st.file_uploader("Upload CV/Resume", type=["pdf", "png", "jpg", "jpeg"])
|
| 225 |
+
cv_resume_text = extract_text(cv_resume_file)
|
| 226 |
+
|
| 227 |
# Tab Layout
|
| 228 |
tab1, tab2, tab3, tab4 = st.tabs(["Cold Email", "Cover Letter", "Research Statement", "SOP"])
|
| 229 |
|
|
|
|
| 237 |
if job_opening_text and cv_resume_text:
|
| 238 |
with st.spinner("Generating..."):
|
| 239 |
try:
|
| 240 |
+
generated_email = chains['email'].run({
|
| 241 |
"professor_name": professor_name,
|
| 242 |
"university_name": university_name,
|
| 243 |
"research_interests": research_interests,
|
| 244 |
"reason": reason,
|
| 245 |
"resume_text": cv_resume_text
|
| 246 |
})
|
| 247 |
+
st.session_state.generated_content['email'] = clean_email_output(generated_email)
|
| 248 |
except Exception as e:
|
| 249 |
st.error(f"Generation error: {e}")
|
| 250 |
else:
|
|
|
|
| 253 |
if st.session_state.generated_content['email']:
|
| 254 |
st.markdown('<div class="output-container">', unsafe_allow_html=True)
|
| 255 |
st.markdown(st.session_state.generated_content['email'])
|
| 256 |
+
st.download_button("Download Email", st.session_state.generated_content['email'],
|
| 257 |
file_name="cold_email.txt", key="email_download")
|
| 258 |
st.markdown('</div>', unsafe_allow_html=True)
|
| 259 |
|
|
|
|
| 267 |
if job_opening_text and cv_resume_text:
|
| 268 |
with st.spinner("Generating..."):
|
| 269 |
try:
|
| 270 |
+
generated_letter = chains['cover_letter'].run({
|
| 271 |
"job_title": job_title,
|
| 272 |
"company": company_name,
|
| 273 |
"key_skills": key_skills,
|
| 274 |
"resume_text": cv_resume_text
|
| 275 |
})
|
| 276 |
+
st.session_state.generated_content['cover_letter'] = clean_cover_letter_output(generated_letter)
|
| 277 |
except Exception as e:
|
| 278 |
st.error(f"Generation error: {e}")
|
| 279 |
else:
|
|
|
|
| 295 |
if st.button("Generate Research Statement", key="research_stmt_btn"):
|
| 296 |
with st.spinner("Generating..."):
|
| 297 |
try:
|
| 298 |
+
generated_statement = chains['research_statement'].run({
|
| 299 |
"research_background": research_background,
|
| 300 |
"key_projects": key_projects,
|
| 301 |
"future_goals": future_goals
|
| 302 |
})
|
| 303 |
+
st.session_state.generated_content['research_statement'] = clean_research_statement_output(generated_statement)
|
| 304 |
except Exception as e:
|
| 305 |
st.error(f"Generation error: {e}")
|
| 306 |
|
|
|
|
| 322 |
if st.button("Generate SOP", key="sop_btn"):
|
| 323 |
with st.spinner("Generating..."):
|
| 324 |
try:
|
| 325 |
+
generated_sop = chains['sop'].run({
|
| 326 |
"motivation": motivation,
|
| 327 |
"academic_background": academic_background,
|
| 328 |
"research_experiences": research_experiences,
|
| 329 |
"career_goals": career_goals,
|
| 330 |
"why_this_program": why_this_program
|
| 331 |
})
|
| 332 |
+
st.session_state.generated_content['sop'] = clean_sop_output(generated_sop)
|
| 333 |
except Exception as e:
|
| 334 |
st.error(f"Generation error: {e}")
|
| 335 |
|