Spaces:
Sleeping
Sleeping
Upload 2 files
Browse files- app.py +138 -0
- requirements.txt +6 -0
app.py
ADDED
|
@@ -0,0 +1,138 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import gradio as gr
|
| 3 |
+
from openai import OpenAI
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
def demo(project_TCGA, output_language="Chinese"):
|
| 7 |
+
name_English, name_Chinese = project_name_TCGA[project_TCGA]
|
| 8 |
+
output1, output2 = None, None
|
| 9 |
+
|
| 10 |
+
if output_language == "Chinese":
|
| 11 |
+
output1 = f"✍️ 简称:{project_TCGA}\n❤️ 中文全称:{name_Chinese}\n💛 英文全称:{name_English}"
|
| 12 |
+
system_instruction = f"您是系统生物学、流行病学、癌症研究和精准医学领域的专家,对{name_Chinese}有着深刻的洞察。"
|
| 13 |
+
prompt_template = f"""
|
| 14 |
+
您的任务是深入分析{name_Chinese}这一复杂疾病,并深入探索其发病和进展中涉及的关键的分子机制与信号通路。让我们一步一步地思考。
|
| 15 |
+
""".strip()
|
| 16 |
+
else:
|
| 17 |
+
output1 = f"✍️ Abbreviation: {project_TCGA}\n❤️ Full name in Chinese: {name_Chinese}\n💛 Full Name in English: {name_English}"
|
| 18 |
+
system_instruction = f"You are an expert in the fields of systems biology, epidemiology, cancer research, and precision medicine, with deep insights into {name_English}."
|
| 19 |
+
prompt_template = f"""
|
| 20 |
+
Your task is to analyze in depth the complex disease of {name_English} and to explore in depth the key molecular mechanisms and signaling pathways involved in its initiation and progression. Let's think step by step.
|
| 21 |
+
""".strip()
|
| 22 |
+
|
| 23 |
+
try:
|
| 24 |
+
# 要实例化一个 OpenAI 对象,你需要设置 OpenAI API Key、Base URL、最大重试次数以及超时限制时间。
|
| 25 |
+
client = OpenAI(
|
| 26 |
+
api_key=os.environ["OPENAI_API_KEY"],
|
| 27 |
+
base_url=os.environ["API_BASE"],
|
| 28 |
+
max_retries=3,
|
| 29 |
+
timeout=60,
|
| 30 |
+
)
|
| 31 |
+
# 调用 client.chat.completions.create,设置关键参数。
|
| 32 |
+
chat_completion = client.chat.completions.create(
|
| 33 |
+
model="gpt-4o-mini-2024-07-18",
|
| 34 |
+
messages=[
|
| 35 |
+
{"role": "system", "content": system_instruction},
|
| 36 |
+
{"role": "user", "content": prompt_template},
|
| 37 |
+
],
|
| 38 |
+
n=1,
|
| 39 |
+
temperature=0.30,
|
| 40 |
+
seed=42,
|
| 41 |
+
max_tokens=2048,
|
| 42 |
+
logprobs=True,
|
| 43 |
+
top_logprobs=3,
|
| 44 |
+
presence_penalty=0,
|
| 45 |
+
frequency_penalty=0,
|
| 46 |
+
)
|
| 47 |
+
resp_text = chat_completion.choices[0].message.content.strip()
|
| 48 |
+
|
| 49 |
+
# 在普通文本框不能用 "**" 渲染加粗,Markdown 才可以。因此,将输入字符串中所有的 "**" 替换为 ""。
|
| 50 |
+
if "**" in resp_text:
|
| 51 |
+
output2 = resp_text.replace("**", "")
|
| 52 |
+
|
| 53 |
+
except Exception as e:
|
| 54 |
+
print(str(e), "Response Error")
|
| 55 |
+
return str(e), "Response Error"
|
| 56 |
+
|
| 57 |
+
return output1, output2
|
| 58 |
+
|
| 59 |
+
|
| 60 |
+
project_name_TCGA = {
|
| 61 |
+
"TCGA-ACC": ["adrenocortical carcinoma", "肾上腺皮质癌"],
|
| 62 |
+
"TCGA-BLCA": ["bladder urothelial carcinoma", "膀胱尿路上皮癌"],
|
| 63 |
+
"TCGA-BRCA": ["breast invasive carcinoma", "浸润性乳腺癌"],
|
| 64 |
+
"TCGA-CESC": [
|
| 65 |
+
"cervical squamous cell carcinoma and endocervical adenocarcinoma",
|
| 66 |
+
"宫颈鳞状细胞癌与宫颈内膜腺癌",
|
| 67 |
+
],
|
| 68 |
+
"TCGA-CHOL": ["cholangiocarcinoma", "胆管癌"],
|
| 69 |
+
"TCGA-COAD": ["colon adenocarcinoma", "结肠腺癌"],
|
| 70 |
+
"TCGA-DLBC": [
|
| 71 |
+
"lymphoid neoplasm diffuse large B-cell lymphoma",
|
| 72 |
+
"弥漫性大 B 细胞淋巴瘤",
|
| 73 |
+
],
|
| 74 |
+
"TCGA-ESCA": ["esophageal carcinoma", "食道癌"],
|
| 75 |
+
"TCGA-GBM": ["glioblastoma multiforme", "多形性胶质母细胞瘤"],
|
| 76 |
+
"TCGA-HNSC": ["head and neck squamous cell carcinoma", "头颈部鳞状细胞癌"],
|
| 77 |
+
"TCGA-KICH": ["kidney chromophobe", "肾嫌色细胞癌"],
|
| 78 |
+
"TCGA-KIRC": ["kidney renal clear cell carcinoma", "肾透明细胞癌"],
|
| 79 |
+
"TCGA-KIRP": ["kidney renal papillary cell carcinoma", "乳头状肾细胞癌"],
|
| 80 |
+
"TCGA-LAML": ["acute myeloid leukemia", "急性髓系白血病"],
|
| 81 |
+
"TCGA-LGG": ["brain lower grade glioma", "低级别脑胶质瘤"],
|
| 82 |
+
"TCGA-LIHC": ["liver hepatocellular carcinoma", "肝细胞癌"],
|
| 83 |
+
"TCGA-LUAD": ["lung adenocarcinoma", "肺腺癌"],
|
| 84 |
+
"TCGA-LUSC": ["lung squamous cell carcinoma", "肺鳞状细胞癌"],
|
| 85 |
+
"TCGA-MESO": ["mesothelioma", "间皮瘤"],
|
| 86 |
+
"TCGA-OV": ["ovarian serous cystadenocarcinoma", "卵巢浆液性囊腺癌"],
|
| 87 |
+
"TCGA-PAAD": ["pancreatic adenocarcinoma", "胰腺腺癌"],
|
| 88 |
+
"TCGA-PCPG": ["pheochromocytoma and paraganglioma", "嗜铬细胞瘤和副神经节瘤"],
|
| 89 |
+
"TCGA-PRAD": ["prostate adenocarcinoma", "前列腺腺癌"],
|
| 90 |
+
"TCGA-READ": ["rectum adenocarcinoma", "直肠腺癌"],
|
| 91 |
+
"TCGA-SARC": ["sarcoma", "肉瘤"],
|
| 92 |
+
"TCGA-SKCM": ["skin cutaneous melanoma", "皮肤黑色素瘤"],
|
| 93 |
+
"TCGA-STAD": ["stomach adenocarcinoma", "胃腺癌"],
|
| 94 |
+
"TCGA-TGCT": ["testicular germ cell tumors", "睾丸生殖细胞肿瘤"],
|
| 95 |
+
"TCGA-THCA": ["thyroid carcinoma", "甲状腺癌"],
|
| 96 |
+
"TCGA-THYM": ["thymoma", "胸腺瘤"],
|
| 97 |
+
"TCGA-UCEC": ["uterine corpus endometrial carcinoma", "子宫体子宫内膜癌"],
|
| 98 |
+
"TCGA-UCS": ["uterine carcinosarcoma", "子宫癌肉瘤"],
|
| 99 |
+
"TCGA-UVM": ["uveal melanoma", "眼内(葡萄膜)黑色素瘤"],
|
| 100 |
+
}
|
| 101 |
+
# print(len(project_name_TCGA.keys()))
|
| 102 |
+
# input_query = input("请输入您要查询的 TCGA 项目名称:")
|
| 103 |
+
# print(project_name_TCGA[input_query])
|
| 104 |
+
print([k for k in project_name_TCGA.keys()])
|
| 105 |
+
|
| 106 |
+
# 支持 Markdown 和 HTML 内容格式:
|
| 107 |
+
# Abbreviations, Full Names and Descriptions of All Cancer Types Covered by TCGA Project.
|
| 108 |
+
desc = """<h1 align="center" style="font-family: Latin Modern Math, sans-serif; font-size: 22px; color: #00FF7F;">🎉 TCGA 项目涵盖的所有癌症类型的缩写、全称和描述 🧬</h1>"""
|
| 109 |
+
outputs = [
|
| 110 |
+
gr.Textbox(
|
| 111 |
+
label="🔎 1. 所查询的癌症类型的全称", show_copy_button=True
|
| 112 |
+
), # 1. The Full Name of The Cancer Type Queried.
|
| 113 |
+
gr.Textbox(
|
| 114 |
+
label="👩⚕️ 2. 深入理解所查询的癌症类型",
|
| 115 |
+
show_copy_button=True,
|
| 116 |
+
), # 2. Insight Into The Cancer Type Being Queried.
|
| 117 |
+
]
|
| 118 |
+
my_demo = gr.Interface(
|
| 119 |
+
fn=demo,
|
| 120 |
+
inputs=[
|
| 121 |
+
gr.Dropdown(
|
| 122 |
+
choices=[k for k in project_name_TCGA.keys()],
|
| 123 |
+
value="TCGA-LUAD",
|
| 124 |
+
allow_custom_value=False,
|
| 125 |
+
label="⌨️ 请输入您要查询的 TCGA 项目名称,例如 TCGA-LUAD。",
|
| 126 |
+
), # Please enter the name of the TCGA project you want to query, such as TCGA-LUAD.
|
| 127 |
+
gr.Dropdown(
|
| 128 |
+
choices=["Chinese", "English"],
|
| 129 |
+
value="Chinese",
|
| 130 |
+
allow_custom_value=False,
|
| 131 |
+
label="👨💻 输出语言",
|
| 132 |
+
),
|
| 133 |
+
],
|
| 134 |
+
outputs=outputs,
|
| 135 |
+
description=desc,
|
| 136 |
+
theme="JohnSmith9982/small_and_pretty",
|
| 137 |
+
)
|
| 138 |
+
my_demo.launch(show_api=False, show_error=True)
|
requirements.txt
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
openai==1.52.2
|
| 2 |
+
requests==2.29.0
|
| 3 |
+
urllib3==1.25.11
|
| 4 |
+
gradio==4.19.2
|
| 5 |
+
gradio_client==0.10.1
|
| 6 |
+
tenacity==8.2.3
|