protein_rag / demo.py
ericzhang1122's picture
Upload folder using huggingface_hub
5c20520 verified
import os
import json
import sys
import tempfile
import gradio as gr
from typing import Dict, List, Optional
from pathlib import Path
from Bio import SeqIO
from io import StringIO
# 添加必要的路径
root_path = os.path.dirname(os.path.abspath(__file__))
sys.path.append(root_path)
sys.path.append(os.path.join(root_path, "Models/ProTrek"))
# 导入所需模块
from interproscan import InterproScan
from Bio.Blast.Applications import NcbiblastpCommandline
from utils.utils import extract_interproscan_metrics, get_seqnid, extract_blast_metrics, rename_interproscan_keys
from go_integration_pipeline import GOIntegrationPipeline
from utils.openai_access import call_chatgpt
from utils.prompts import FUNCTION_PROMPT
def get_prompt_template(selected_info_types=None):
"""
获取prompt模板,支持可选的信息类型
Args:
selected_info_types: 需要包含的信息类型列表,如['motif', 'go', 'superfamily', 'panther']
"""
if selected_info_types is None:
selected_info_types = ['motif', 'go'] # 默认包含motif和go信息
PROMPT_TEMPLATE = FUNCTION_PROMPT + '\n' + """
input information:
{%- if 'motif' in selected_info_types and motif_pfam %}
motif:{% for motif_id, motif_info in motif_pfam.items() %}
{{motif_id}}: {{motif_info}}
{% endfor %}
{%- endif %}
{%- if 'go' in selected_info_types and go_data.status == 'success' %}
GO:{% for go_entry in go_data.go_annotations %}
▢ GO term{{loop.index}}: {{go_entry.go_id}}
• definition: {{ go_data.all_related_definitions.get(go_entry.go_id, 'not found definition') }}
{% endfor %}
{%- endif %}
{%- for info_type in selected_info_types %}
{%- if info_type not in ['motif', 'go'] and interpro_descriptions.get(info_type) %}
{{info_type}}:{% for ipr_id, ipr_info in interpro_descriptions[info_type].items() %}
▢ {{ipr_id}}: {{ipr_info.name}}
• description: {{ipr_info.abstract}}
{% endfor %}
{%- endif %}
{%- endfor %}
question: \n {{question}}
"""
return PROMPT_TEMPLATE
class ProteinAnalysisDemo:
def __init__(self):
"""
蛋白质分析演示类
"""
self.blast_database = "uniprot_swissprot"
self.expect_value = 0.01
self.interproscan_path = "interproscan/interproscan-5.75-106.0/interproscan.sh"
self.interproscan_libraries = [
"PFAM", "PIRSR", "PROSITE_PROFILES", "SUPERFAMILY", "PRINTS",
"PANTHER", "CDD", "GENE3D", "NCBIFAM", "SFLM", "MOBIDB_LITE",
"COILS", "PROSITE_PATTERNS", "FUNFAM", "SMART"
]
self.go_topk = 2
self.selected_info_types = ['motif', 'go']
# 文件路径配置
self.pfam_descriptions_path = 'data/raw_data/all_pfam_descriptions.json'
self.go_info_path = 'data/raw_data/go.json'
self.interpro_data_path = 'data/raw_data/interpro_data.json'
# 初始化GO整合管道
self.go_pipeline = GOIntegrationPipeline(topk=self.go_topk)
# 初始化InterPro管理器(如果需要)
self.interpro_manager = None
other_types = [t for t in self.selected_info_types if t not in ['motif', 'go']]
if other_types and os.path.exists(self.interpro_data_path):
try:
from utils.generate_protein_prompt import get_interpro_manager
self.interpro_manager = get_interpro_manager(self.interpro_data_path, None)
except Exception as e:
print(f"初始化InterPro管理器失败: {str(e)}")
def validate_protein_sequence(self, sequence: str) -> bool:
"""
验证蛋白质序列格式
"""
if not sequence:
return False
# 移除空白字符
sequence = sequence.strip().upper()
# 检查是否包含有效的氨基酸字符
valid_aa = set('ACDEFGHIKLMNPQRSTVWY')
sequence_chars = set(sequence.replace('\n', '').replace(' ', ''))
return sequence_chars.issubset(valid_aa) and len(sequence) > 0
def parse_fasta_content(self, fasta_content: str) -> tuple:
"""
解析FASTA内容,返回第一个序列
"""
try:
fasta_io = StringIO(fasta_content)
records = list(SeqIO.parse(fasta_io, "fasta"))
if not records:
return None, "FASTA文件中没有找到有效的序列"
if len(records) > 1:
return None, "演示版本只支持单一序列,检测到多个序列"
record = records[0]
return str(record.seq), f"成功解析序列 ID: {record.id}"
except Exception as e:
return None, f"解析FASTA文件出错: {str(e)}"
def create_temp_fasta(self, sequence: str, seq_id: str = "demo_protein") -> str:
"""
创建临时FASTA文件
"""
temp_file = tempfile.NamedTemporaryFile(mode='w', suffix='.fasta', delete=False)
temp_file.write(f">{seq_id}\n{sequence}\n")
temp_file.close()
return temp_file.name
def run_blast_analysis(self, fasta_file: str, temp_dir: str) -> Dict:
"""
运行BLAST分析
"""
blast_xml = os.path.join(temp_dir, "blast_results.xml")
try:
blast_cmd = NcbiblastpCommandline(
query=fasta_file,
db=self.blast_database,
out=blast_xml,
outfmt=5, # XML格式
evalue=self.expect_value
)
blast_cmd()
# 提取BLAST结果
blast_results = extract_blast_metrics(blast_xml)
# 获取序列字典
seq_dict = get_seqnid(fasta_file)
blast_info = {}
for uid, info in blast_results.items():
blast_info[uid] = {"sequence": seq_dict[uid], "blast_results": info}
return blast_info
except Exception as e:
print(f"BLAST分析出错: {str(e)}")
return {}
finally:
if os.path.exists(blast_xml):
os.remove(blast_xml)
def run_interproscan_analysis(self, fasta_file: str, temp_dir: str) -> Dict:
"""
运行InterProScan分析
"""
interproscan_json = os.path.join(temp_dir, "interproscan_results.json")
try:
interproscan = InterproScan(self.interproscan_path)
input_args = {
"fasta_file": fasta_file,
"goterms": True,
"pathways": True,
"save_dir": interproscan_json
}
interproscan.run(**input_args)
# 提取InterProScan结果
interproscan_results = extract_interproscan_metrics(
interproscan_json,
librarys=self.interproscan_libraries
)
# 获取序列字典
seq_dict = get_seqnid(fasta_file)
interproscan_info = {}
for id, seq in seq_dict.items():
info = interproscan_results[seq]
info = rename_interproscan_keys(info)
interproscan_info[id] = {"sequence": seq, "interproscan_results": info}
return interproscan_info
except Exception as e:
print(f"InterProScan分析出错: {str(e)}")
return {}
finally:
if os.path.exists(interproscan_json):
os.remove(interproscan_json)
def generate_prompt(self, protein_id: str, interproscan_info: Dict,
protein_go_dict: Dict, question: str) -> str:
"""
从内存中的数据生成prompt,包含完整的motif和GO定义
"""
try:
from utils.protein_go_analysis import get_go_definition
from jinja2 import Template
# from utils.generate_protein_prompt import get_prompt_template
# 获取GO分析结果
go_ids = protein_go_dict.get(protein_id, [])
go_annotations = []
all_related_definitions = {}
if go_ids:
for go_id in go_ids:
# 确保GO ID格式正确
clean_go_id = go_id.split(":")[-1] if ":" in go_id else go_id
go_annotations.append({"go_id": clean_go_id})
# 获取GO定义
if os.path.exists(self.go_info_path):
definition = get_go_definition(clean_go_id, self.go_info_path)
if definition:
all_related_definitions[clean_go_id] = definition
# 获取motif信息
motif_pfam = {}
if os.path.exists(self.pfam_descriptions_path):
try:
# 从interproscan结果中提取pfam信息
interproscan_results = interproscan_info[protein_id].get('interproscan_results', {})
pfam_entries = interproscan_results.get('pfam_id', [])
# 加载pfam描述
with open(self.pfam_descriptions_path, 'r') as f:
pfam_descriptions = json.load(f)
# 构建motif_pfam字典
for entry in pfam_entries:
for pfam_id, ipr_id in entry.items():
if pfam_id and pfam_id in pfam_descriptions:
motif_pfam[pfam_id] = pfam_descriptions[pfam_id]['description']
except Exception as e:
print(f"获取motif信息时出错: {str(e)}")
# 获取InterPro描述信息
interpro_descriptions = {}
other_types = [t for t in self.selected_info_types if t not in ['motif', 'go']]
if other_types and self.interpro_manager:
interpro_descriptions = self.interpro_manager.get_description(protein_id, other_types)
# 准备模板数据
template_data = {
"protein_id": protein_id,
"selected_info_types": self.selected_info_types,
"go_data": {
"status": "success" if go_annotations else "no_data",
"go_annotations": go_annotations,
"all_related_definitions": all_related_definitions
},
"motif_pfam": motif_pfam,
"interpro_descriptions": interpro_descriptions,
"question": question
}
# 使用模板生成prompt
PROMPT_TEMPLATE = get_prompt_template(self.selected_info_types) # demo版本不使用lmdb
template = Template(PROMPT_TEMPLATE)
return template.render(**template_data)
except Exception as e:
print(f"生成prompt时出错 (protein_id: {protein_id}): {str(e)}")
# 如果出错,返回简化版本的prompt
return self._generate_fallback_prompt(protein_id, interproscan_info, protein_go_dict, question)
def _generate_fallback_prompt(self, protein_id: str, interproscan_info: Dict,
protein_go_dict: Dict, question: str) -> str:
"""
生成备用prompt(当主要方法失败时使用)
"""
from utils.prompts import FUNCTION_PROMPT
prompt_parts = [FUNCTION_PROMPT]
prompt_parts.append("\ninput information:")
# 添加motif信息
if 'motif' in self.selected_info_types:
interproscan_results = interproscan_info[protein_id].get('interproscan_results', {})
pfam_entries = interproscan_results.get('pfam_id', [])
if pfam_entries:
prompt_parts.append("\nmotif:")
for entry in pfam_entries:
for key, value in entry.items():
if value:
prompt_parts.append(f"{value}: motif information")
# 添加GO信息
if 'go' in self.selected_info_types:
go_ids = protein_go_dict.get(protein_id, [])
if go_ids:
prompt_parts.append("\nGO:")
for i, go_id in enumerate(go_ids[:10], 1):
prompt_parts.append(f"▢ GO term{i}: {go_id}")
prompt_parts.append(f"• definition: GO term definition")
# 添加用户问题
prompt_parts.append(f"\nquestion: \n{question}")
return "\n".join(prompt_parts)
def analyze_protein(self, sequence_input: str, fasta_file, question: str) -> str:
"""
分析蛋白质序列并回答问题
"""
if not question.strip():
return "请输入您的问题"
# 确定使用哪个序列输入
final_sequence = None
sequence_source = ""
if fasta_file is not None:
# 优先使用上传的文件
try:
fasta_content = fasta_file.read().decode('utf-8')
final_sequence, parse_msg = self.parse_fasta_content(fasta_content)
if final_sequence is None:
return f"文件解析错误: {parse_msg}"
sequence_source = f"来自上传文件: {parse_msg}"
except Exception as e:
return f"读取上传文件出错: {str(e)}"
elif sequence_input.strip():
# 使用文本框输入的序列
if self.validate_protein_sequence(sequence_input):
final_sequence = sequence_input.strip().upper().replace('\n', '').replace(' ', '')
sequence_source = "来自文本框输入"
else:
return "输入的序列格式不正确,请输入有效的蛋白质序列"
else:
return "请输入蛋白质序列或上传FASTA文件"
# 创建临时目录和文件
with tempfile.TemporaryDirectory() as temp_dir:
try:
# 创建临时FASTA文件
temp_fasta = self.create_temp_fasta(final_sequence, "demo_protein")
# 运行分析
status_msg = f"序列来源: {sequence_source}\n序列长度: {len(final_sequence)} 氨基酸\n\n正在进行分析...\n"
# 步骤1: BLAST和InterProScan分析
status_msg += "步骤1: 运行BLAST分析...\n"
blast_info = self.run_blast_analysis(temp_fasta, temp_dir)
status_msg += "步骤2: 运行InterProScan分析...\n"
interproscan_info = self.run_interproscan_analysis(temp_fasta, temp_dir)
if not blast_info or not interproscan_info:
return status_msg + "分析失败: 无法获取BLAST或InterProScan结果"
# 步骤2: 整合GO信息
status_msg += "步骤3: 整合GO信息...\n"
protein_go_dict = self.go_pipeline.first_level_filtering(interproscan_info, blast_info)
# 步骤3: 生成prompt
status_msg += "步骤4: 生成分析prompt...\n"
protein_id = "demo_protein"
prompt = self.generate_prompt(protein_id, interproscan_info, protein_go_dict, question)
# 步骤4: 调用LLM生成答案
status_msg += "步骤5: 生成答案...\n"
llm_response = call_chatgpt(prompt)
# 组织最终结果
result = f"""
{status_msg}
=== 分析完成 ===
问题: {question}
答案: {llm_response}
=== 分析详情 ===
- BLAST匹配数: {len(blast_info.get(protein_id, {}).get('blast_results', []))}
- InterProScan域数: {len(interproscan_info.get(protein_id, {}).get('interproscan_results', {}).get('pfam_id', []))}
- GO术语数: {len(protein_go_dict.get(protein_id, []))}
"""
return result
except Exception as e:
return f"分析过程中出错: {str(e)}"
finally:
# 清理临时文件
if 'temp_fasta' in locals() and os.path.exists(temp_fasta):
os.remove(temp_fasta)
def create_demo():
"""
创建Gradio演示界面
"""
analyzer = ProteinAnalysisDemo()
with gr.Blocks(title="蛋白质功能分析演示") as demo:
gr.Markdown("# 🧬 蛋白质功能分析演示")
gr.Markdown("输入蛋白质序列和问题,AI将基于BLAST、InterProScan和GO信息为您提供专业分析")
with gr.Row():
with gr.Column(scale=1):
gr.Markdown("### 📝 序列输入")
sequence_input = gr.Textbox(
label="蛋白质序列",
placeholder="请输入蛋白质序列(单字母氨基酸代码)...",
lines=5,
max_lines=10
)
gr.Markdown("**或者**")
fasta_file = gr.File(
label="上传FASTA文件",
file_types=[".fasta", ".fa", ".fas"],
file_count="single"
)
gr.Markdown("### ❓ 您的问题")
question_input = gr.Textbox(
label="问题",
placeholder="请输入关于该蛋白质的问题,例如:这个蛋白质的主要功能是什么?",
lines=3
)
analyze_btn = gr.Button("🔍 开始分析", variant="primary", size="lg")
with gr.Column(scale=2):
gr.Markdown("### 📊 分析结果")
output = gr.Textbox(
label="分析结果",
lines=20,
max_lines=30,
show_copy_button=True
)
# 示例
gr.Markdown("### 💡 示例")
gr.Examples(
examples=[
["MKALIVLGLVLLSVTVQGKVFERCELARTLKRLGMDGYRGISLANWMCLAKWESGYNTRATNYNAGDRSTDYGIFQINSRYWCNDGKTPGAVNACHLSCSALLQDNIADAVACAKRVVRDPQGIRAWVAWRNRCQNRDVRQYVQGCGV", "这个蛋白质的主要功能是什么?"],
["MGSSHHHHHHSSGLVPRGSHMRGPNPTAASLEASAGPFTVRSFTVSRPSGYGAGTVYYPTNAGGTVGAIAIVPGYTARQSSIKWWGPRLASHGFVVITIDTNSTLDQPSSRSSQQMAALRQVASLNGTSSSPIYGKVDTARMGVMGWSMGGGGSLISAANNPSLKAAAPQAPWDSSTNFSSVTVPTLIFACENDSIAPVNSSALPIYDSMSRNAKQFLEINGGSHSCANSGNSNQALIGKKGVAWMKRFPTSREJ", "这个蛋白质可能参与哪些生物学过程?"],
["ATGAGTGAACGTCTGAAATCTATCATCACCGTCGACGACGAGAACGTCAAGCTGATCGACAAGATCCTGGCCTCCATCAAGGACCTGAACGAGCTGGTGGACATGATCGACGAGATCAAGAACGTCGACGACGAGCTGATCGACAAGATCCTGGCC", "这个序列编码的蛋白质具有什么结构特征?"]
],
inputs=[sequence_input, question_input]
)
analyze_btn.click(
fn=analyzer.analyze_protein,
inputs=[sequence_input, fasta_file, question_input],
outputs=[output]
)
return demo
if __name__ == "__main__":
demo = create_demo()
demo.launch(
server_name="0.0.0.0",
server_port=30002,
share=True,
debug=False
)