Update Dockerfile

Dockerfile

@@ -39,39 +39,48 @@ RUN pip3 install --no-cache-dir transvar
 # 创建数据目录
 RUN mkdir -p /data/transvar_db/refseq_hg38 /data/transvar_db/refseq_hg19
 
-# ========== 下载数据脚本 ==========
-# 如果设置了 HF_DATASET，从数据集下载
-RUN if [ -n "$HF_DATASET" ]; then \
-    echo "Downloading from HF Dataset: $HF_DATASET" && \
-    python3 -c "import os; from huggingface_hub import snapshot_download; dataset=os.environ.get('HF_DATASET',''); local=snapshot_download(repo_id=dataset, repo_type='dataset', cache_dir='/tmp/hf_cache'); os.system(f'cp -r {local}/* /data/transvar_db/')" || true; \
-    fi
-
-# 如果没有设置 HF_DATASET，从 UCSC 下载（首次构建）
+# ========== 下载参考基因组数据 ==========
 WORKDIR /data/transvar_db/refseq_hg38
 
-RUN if [ ! -f hg38.fa ] || [ ! -f ncbiRefSeq.txt.gz ]; then \
-    echo "Downloading hg38 from UCSC..."; \
+# 下载 hg38 参考基因组
+RUN if [ ! -f hg38.fa ]; then \
     wget -q -O hg38.fa.gz https://hgdownload.soe.ucsc.edu/goldenPath/hg38/bigZips/hg38.fa.gz && gunzip -f hg38.fa.gz && samtools faidx hg38.fa; \
+    fi
+
+# 下载 hg38 RefSeq
+RUN if [ ! -f ncbiRefSeq.txt.gz ]; then \
     wget -q -O ncbiRefSeq.txt.gz https://hgdownload.soe.ucsc.edu/goldenPath/hg38/database/ncbiRefSeq.txt.gz; \
-    transvar index --refseq ncbiRefSeq.txt.gz; \
-    transvar config -k reference -v /data/transvar_db/refseq_hg38/hg38.fa --refversion hg38_refseq; \
-    transvar config -k refseq -v /data/transvar_db/refseq_hg38/ncbiRefSeq.txt.gz --refversion hg38_refseq; \
     fi
 
+# 下载 hg19 参考基因组
 WORKDIR /data/transvar_db/refseq_hg19
 
-RUN if [ ! -f hg19.fa ] || [ ! -f ncbiRefSeq.txt.gz ]; then \
-    echo "Downloading hg19 from UCSC..."; \
+RUN if [ ! -f hg19.fa ]; then \
     wget -q -O hg19.fa.gz https://hgdownload.soe.ucsc.edu/goldenPath/hg19/bigZips/hg19.fa.gz && gunzip -f hg19.fa.gz && samtools faidx hg19.fa; \
+    fi
+
+# 下载 hg19 RefSeq
+RUN if [ ! -f ncbiRefSeq.txt.gz ]; then \
     wget -q -O ncbiRefSeq.txt.gz https://hgdownload.soe.ucsc.edu/goldenPath/hg19/database/ncbiRefSeq.txt.gz; \
-    transvar index --refseq ncbiRefSeq.txt.gz; \
-    transvar config -k reference -v /data/transvar_db/refseq_hg19/hg19.fa --refversion hg19_refseq; \
-    transvar config -k refseq -v /data/transvar_db/refseq_hg19/ncbiRefSeq.txt.gz --refversion hg19_refseq; \
     fi
 
-# 返回工作目录
+# ========== 创建启动脚本 ==========
 WORKDIR /app
+RUN printf '#!/bin/bash\n\
+echo "Setting up transvar database..."\n\
+if [ ! -f /data/transvar_db/refseq_hg38/ncbiRefSeq.txt ]; then\n\
+    gunzip -c /data/transvar_db/refseq_hg38/ncbiRefSeq.txt.gz > /data/transvar_db/refseq_hg38/ncbiRefSeq.txt\n\
+fi\n\
+transvar config -k reference -v /data/transvar_db/refseq_hg38/hg38.fa --refversion hg38_refseq\n\
+transvar config -k refseq -v /data/transvar_db/refseq_hg38/ncbiRefSeq.txt --refversion hg38_refseq\n\
+if [ ! -f /data/transvar_db/refseq_hg19/ncbiRefSeq.txt ]; then\n\
+    gunzip -c /data/transvar_db/refseq_hg19/ncbiRefSeq.txt.gz > /data/transvar_db/refseq_hg19/ncbiRefSeq.txt\n\
+fi\n\
+transvar config -k reference -v /data/transvar_db/refseq_hg19/hg19.fa --refversion hg19_refseq\n\
+transvar config -k refseq -v /data/transvar_db/refseq_hg19/ncbiRefSeq.txt --refversion hg19_refseq\n\
+echo "Starting server..."\n\
+exec python3 -c "import os; import uvicorn; from server import app; uvicorn.run(app, host=\0470.0.0.0\047, port=int(os.environ.get(\047PORT\047, 7860)))"\n' > /app/start.sh && chmod +x /app/start.sh
 
 EXPOSE 7860
 
-CMD python3 -c "import os; import uvicorn; from server import app; uvicorn.run(app, host='0.0.0.0', port=int(os.environ.get('PORT', 7860)))"
+CMD ["/app/start.sh"]