czty commited on
Commit
96d6cfe
·
verified ·
1 Parent(s): 0d8232c

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_anarci/Dockerfile +40 -0
  2. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_anarci/app/requirements.txt +1 -0
  3. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_anarci/requirements.txt +2 -0
  4. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_ariba/app/ariba_server.py +509 -0
  5. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_ariba/app/ariba_shim_server.py +45 -0
  6. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/Dockerfile +40 -0
  7. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/app/arvados-python-client_server.py +298 -0
  8. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/app/arvados-python-client_shim_server.py +45 -0
  9. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/app/requirements.txt +1 -0
  10. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/docker-compose.yml +22 -0
  11. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/environment.yaml +10 -0
  12. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/requirements.txt +2 -0
  13. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_augur/requirements.txt +2 -0
  14. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/Dockerfile +40 -0
  15. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/app/bcbio-gff_server.py +192 -0
  16. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/app/bcbio-gff_shim_server.py +45 -0
  17. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/app/requirements.txt +1 -0
  18. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/docker-compose.yml +22 -0
  19. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/environment.yaml +10 -0
  20. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/requirements.txt +2 -0
  21. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/Dockerfile +40 -0
  22. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/app/__pycache__/bioblend_server.cpython-310.pyc +0 -0
  23. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/app/bioblend_server.py +111 -0
  24. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/app/bioblend_shim_server.py +45 -0
  25. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/app/requirements.txt +1 -0
  26. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/docker-compose.yml +22 -0
  27. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/environment.yaml +10 -0
  28. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/requirements.txt +3 -0
  29. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/Dockerfile +40 -0
  30. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/app/bioconductor-affy_server.py +196 -0
  31. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/app/bioconductor-affy_shim_server.py +45 -0
  32. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/app/requirements.txt +1 -0
  33. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/docker-compose.yml +22 -0
  34. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/environment.yaml +10 -0
  35. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/requirements.txt +2 -0
  36. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/Dockerfile +40 -0
  37. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/app/bioconductor-alabaster.sfe_server.py +213 -0
  38. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/app/bioconductor-alabaster.sfe_shim_server.py +45 -0
  39. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/app/requirements.txt +1 -0
  40. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/docker-compose.yml +22 -0
  41. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/environment.yaml +10 -0
  42. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/requirements.txt +2 -0
  43. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/Dockerfile +40 -0
  44. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/app/bioconductor-biovizbase_server.py +437 -0
  45. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/app/bioconductor-biovizbase_shim_server.py +45 -0
  46. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/app/requirements.txt +1 -0
  47. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/docker-compose.yml +22 -0
  48. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/environment.yaml +10 -0
  49. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/requirements.txt +2 -0
  50. BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-bulksignalr/Dockerfile +40 -0
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_anarci/Dockerfile ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ FROM python:3.10-slim
3
+
4
+ # Install system dependencies
5
+ RUN apt-get update && apt-get install -y default-jre wget curl && apt-get clean && rm -rf /var/lib/apt/lists/*
6
+
7
+ # Install Miniconda
8
+ RUN wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O /tmp/miniconda.sh && bash /tmp/miniconda.sh -b -p /opt/conda && rm /tmp/miniconda.sh
9
+
10
+ # Add conda to PATH
11
+ ENV PATH="/opt/conda/bin:$PATH"
12
+
13
+ # Install anarci via conda (e.g., from bioconda)
14
+ RUN conda install -c bioconda anarci -y && conda clean -a
15
+
16
+ # Install Python dependencies
17
+ RUN pip install uv
18
+ RUN uv pip install --system fastmcp
19
+
20
+ # Create app directory
21
+ WORKDIR /app
22
+
23
+ # Copy your MCP server
24
+ COPY app/anarci_server.py /app/
25
+
26
+ # Create workspace and output directories
27
+ RUN mkdir -p /app/workspace /app/output
28
+
29
+ # Make sure the server script is executable
30
+ RUN chmod +x /app/anarci_server.py
31
+
32
+ # Expose port for MCP over HTTP (optional)
33
+ EXPOSE 8000
34
+
35
+ # Health check
36
+ HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 CMD python -c "import sys; sys.exit(0)"
37
+
38
+ # Default command runs the MCP server via stdio
39
+ CMD ["python", "/app/anarci_server.py"]
40
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_anarci/app/requirements.txt ADDED
@@ -0,0 +1 @@
 
 
1
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_anarci/requirements.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ fastmcp
2
+ mcp
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_ariba/app/ariba_server.py ADDED
@@ -0,0 +1,509 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import subprocess
2
+ import tempfile
3
+ import os
4
+ from pathlib import Path
5
+ from typing import Optional, List, Literal
6
+
7
+ # MCP decorator is not needed for the final code, as per instructions.
8
+ # Assuming it's provided by the MCP framework.
9
+ # from mcp import tool as mcp_tool
10
+ # For linting purposes, we can create a dummy decorator.
11
+ class mcp:
12
+ def tool(func):
13
+ return func
14
+
15
+ @mcp.tool
16
+ def pubmlstget(
17
+ pubmlst_db: str,
18
+ outdir: Path,
19
+ force: bool = False,
20
+ noverify: bool = False,
21
+ ):
22
+ """
23
+ Download data from a pubmlst database.
24
+
25
+ Args:
26
+ pubmlst_db: Name of the pubmlst database to download.
27
+ outdir: Output directory to save the downloaded data.
28
+ force: Overwrite existing files if they exist.
29
+ noverify: Do not verify the SSL certificate.
30
+ """
31
+ # Input validation
32
+ if not pubmlst_db:
33
+ raise ValueError("pubmlst_db must be provided.")
34
+
35
+ try:
36
+ outdir.mkdir(parents=True, exist_ok=True)
37
+ except Exception as e:
38
+ raise IOError(f"Could not create output directory {outdir}: {e}")
39
+
40
+ # Command construction
41
+ cmd = ["ariba", "pubmlstget"]
42
+ if force:
43
+ cmd.append("--force")
44
+ if noverify:
45
+ cmd.append("--noverify")
46
+ cmd.extend([pubmlst_db, str(outdir)])
47
+
48
+ # Subprocess execution
49
+ try:
50
+ result = subprocess.run(
51
+ cmd,
52
+ check=True,
53
+ capture_output=True,
54
+ text=True,
55
+ )
56
+ output_files = [str(p) for p in outdir.rglob("*") if p.is_file()]
57
+ return {
58
+ "command_executed": " ".join(cmd),
59
+ "stdout": result.stdout,
60
+ "stderr": result.stderr,
61
+ "output_files": output_files,
62
+ }
63
+ except FileNotFoundError:
64
+ raise RuntimeError("ariba command not found. Please ensure it is in your PATH.")
65
+ except subprocess.CalledProcessError as e:
66
+ return {
67
+ "command_executed": " ".join(cmd),
68
+ "stdout": e.stdout,
69
+ "stderr": e.stderr,
70
+ "error": "ARIBA pubmlstget failed",
71
+ "return_code": e.returncode,
72
+ }
73
+
74
+ @mcp.tool
75
+ def getref(
76
+ ref_name: Literal[
77
+ "argannot", "card", "megares", "plasmidfinder", "resfinder",
78
+ "srst2_argannot", "vfdb_core", "vfdb_full", "virfinder"
79
+ ],
80
+ outdir: Path,
81
+ force: bool = False,
82
+ metadata_only: bool = False,
83
+ ):
84
+ """
85
+ Download reference data from NCBI/RefSeq or other sources.
86
+
87
+ Args:
88
+ ref_name: Name of reference data to download.
89
+ outdir: Output directory to save the reference data.
90
+ force: Overwrite existing files if they exist.
91
+ metadata_only: Only download metadata, not the sequences.
92
+ """
93
+ # Input validation
94
+ try:
95
+ outdir.mkdir(parents=True, exist_ok=True)
96
+ except Exception as e:
97
+ raise IOError(f"Could not create output directory {outdir}: {e}")
98
+
99
+ # Command construction
100
+ cmd = ["ariba", "getref"]
101
+ if force:
102
+ cmd.append("--force")
103
+ if metadata_only:
104
+ cmd.append("--metadata_only")
105
+ cmd.extend([ref_name, str(outdir)])
106
+
107
+ # Subprocess execution
108
+ try:
109
+ result = subprocess.run(
110
+ cmd,
111
+ check=True,
112
+ capture_output=True,
113
+ text=True,
114
+ )
115
+ output_files = [str(p) for p in outdir.rglob("*") if p.is_file()]
116
+ return {
117
+ "command_executed": " ".join(cmd),
118
+ "stdout": result.stdout,
119
+ "stderr": result.stderr,
120
+ "output_files": output_files,
121
+ }
122
+ except FileNotFoundError:
123
+ raise RuntimeError("ariba command not found. Please ensure it is in your PATH.")
124
+ except subprocess.CalledProcessError as e:
125
+ return {
126
+ "command_executed": " ".join(cmd),
127
+ "stdout": e.stdout,
128
+ "stderr": e.stderr,
129
+ "error": "ARIBA getref failed",
130
+ "return_code": e.returncode,
131
+ }
132
+
133
+ @mcp.tool
134
+ def prepareref(
135
+ fasta_in: Path,
136
+ metadata_in: Path,
137
+ outdir: Path,
138
+ min_length: int = 1,
139
+ max_length: int = 10000,
140
+ max_n: int = 5,
141
+ no_cdhit: bool = False,
142
+ cdhit_clusters: Optional[Path] = None,
143
+ cdhit_threads: int = 1,
144
+ cdhit_mem: int = 400,
145
+ force: bool = False,
146
+ tmp_dir: Optional[Path] = None,
147
+ verbose: bool = False,
148
+ debug: bool = False,
149
+ ):
150
+ """
151
+ Prepare reference data for use with ARIBA.
152
+
153
+ Args:
154
+ fasta_in: Input FASTA file of sequences.
155
+ metadata_in: Input TSV file of metadata.
156
+ outdir: Output directory for prepared reference.
157
+ min_length: Minimum sequence length.
158
+ max_length: Maximum sequence length.
159
+ max_n: Maximum number of Ns allowed in a sequence.
160
+ no_cdhit: Do not run cd-hit.
161
+ cdhit_clusters: File of cd-hit clusters. Implies --no_cdhit.
162
+ cdhit_threads: Threads to use for cd-hit.
163
+ cdhit_mem: Memory for cd-hit in MB.
164
+ force: Overwrite existing files.
165
+ tmp_dir: Use this directory for temporary files.
166
+ verbose: Print STDOUT and STDERR from external programs.
167
+ debug: Print debugging messages.
168
+ """
169
+ # Input validation
170
+ if not fasta_in.is_file():
171
+ raise FileNotFoundError(f"Input FASTA file not found: {fasta_in}")
172
+ if not metadata_in.is_file():
173
+ raise FileNotFoundError(f"Input metadata file not found: {metadata_in}")
174
+ if cdhit_clusters and not cdhit_clusters.is_file():
175
+ raise FileNotFoundError(f"cd-hit clusters file not found: {cdhit_clusters}")
176
+ if min_length <= 0:
177
+ raise ValueError("min_length must be a positive integer.")
178
+ if cdhit_threads <= 0:
179
+ raise ValueError("cdhit_threads must be a positive integer.")
180
+
181
+ try:
182
+ outdir.mkdir(parents=True, exist_ok=True)
183
+ except Exception as e:
184
+ raise IOError(f"Could not create output directory {outdir}: {e}")
185
+
186
+ # Command construction
187
+ cmd = ["ariba", "prepareref",
188
+ "--fasta_in", str(fasta_in),
189
+ "--metadata_in", str(metadata_in),
190
+ "--outdir", str(outdir),
191
+ "--min_length", str(min_length),
192
+ "--max_length", str(max_length),
193
+ "--max_N", str(max_n),
194
+ "--cdhit_threads", str(cdhit_threads),
195
+ "--cdhit_mem", str(cdhit_mem)]
196
+
197
+ if no_cdhit:
198
+ cmd.append("--no_cdhit")
199
+ if cdhit_clusters:
200
+ cmd.extend(["--cdhit_clusters", str(cdhit_clusters)])
201
+ if force:
202
+ cmd.append("--force")
203
+ if tmp_dir:
204
+ cmd.extend(["--tmp_dir", str(tmp_dir)])
205
+ if verbose:
206
+ cmd.append("--verbose")
207
+ if debug:
208
+ cmd.append("--debug")
209
+
210
+ # Subprocess execution
211
+ try:
212
+ result = subprocess.run(
213
+ cmd,
214
+ check=True,
215
+ capture_output=True,
216
+ text=True,
217
+ )
218
+ output_files = [str(p) for p in outdir.rglob("*") if p.is_file()]
219
+ return {
220
+ "command_executed": " ".join(cmd),
221
+ "stdout": result.stdout,
222
+ "stderr": result.stderr,
223
+ "output_files": output_files,
224
+ }
225
+ except FileNotFoundError:
226
+ raise RuntimeError("ariba command not found. Please ensure it is in your PATH.")
227
+ except subprocess.CalledProcessError as e:
228
+ return {
229
+ "command_executed": " ".join(cmd),
230
+ "stdout": e.stdout,
231
+ "stderr": e.stderr,
232
+ "error": "ARIBA prepareref failed",
233
+ "return_code": e.returncode,
234
+ }
235
+
236
+ @mcp.tool
237
+ def run(
238
+ ref_dir: Path,
239
+ reads1: Path,
240
+ reads2: Path,
241
+ outdir: Path,
242
+ assembly_cov: int = 10,
243
+ gene_cov_cutoff: float = 0.95,
244
+ gene_nt_extend: int = 10,
245
+ local_th: float = 0.9,
246
+ min_gene_cov: int = 1,
247
+ assembled_threshold: float = 0.95,
248
+ nucmer_min_id: int = 90,
249
+ nucmer_min_len: int = 50,
250
+ nucmer_breaklen: int = 10,
251
+ nucmer_max_gap: int = 20,
252
+ assembler: Literal["fermilite", "spades"] = "fermilite",
253
+ spades_options: str = "--careful",
254
+ threads: int = 1,
255
+ force: bool = False,
256
+ tmp_dir: Optional[Path] = None,
257
+ verbose: bool = False,
258
+ debug: bool = False,
259
+ noclean: bool = False,
260
+ unique_map_only: bool = False,
261
+ other_ref_fasta: Optional[Path] = None,
262
+ filestem: Optional[str] = None,
263
+ ):
264
+ """
265
+ Run the main ARIBA pipeline to identify genes from paired sequencing reads.
266
+
267
+ Args:
268
+ ref_dir: Reference directory, created by 'prepareref'.
269
+ reads1: First paired-end reads file.
270
+ reads2: Second paired-end reads file.
271
+ outdir: Output directory for results.
272
+ assembly_cov: Minimum coverage of internal contig flank.
273
+ gene_cov_cutoff: Minimum proportion of gene length covered by reads.
274
+ gene_nt_extend: Max nucleotides to extend assembly past gene ends.
275
+ local_th: Minimum proportion of bases in local assembly identical to reference.
276
+ min_gene_cov: Minimum read coverage depth on any base of a gene.
277
+ assembled_threshold: Minimum proportion of bases in a gene covered by local assembly.
278
+ nucmer_min_id: nucmer --minid option.
279
+ nucmer_min_len: nucmer --mincluster option.
280
+ nucmer_breaklen: nucmer --breaklen option.
281
+ nucmer_max_gap: nucmer --maxgap option.
282
+ assembler: Assembler to use.
283
+ spades_options: Options to pass to spades.
284
+ threads: Number of threads to use.
285
+ force: Overwrite existing output files.
286
+ tmp_dir: Use this directory for temporary files.
287
+ verbose: Print STDOUT and STDERR from external programs.
288
+ debug: Print debugging messages.
289
+ noclean: Do not clean up intermediate files.
290
+ unique_map_only: Only use uniquely mapping reads.
291
+ other_ref_fasta: FASTA file of other reference sequences (e.g., host genome).
292
+ filestem: Filestem to use for output files.
293
+ """
294
+ # Input validation
295
+ if not ref_dir.is_dir():
296
+ raise FileNotFoundError(f"Reference directory not found: {ref_dir}")
297
+ if not reads1.is_file():
298
+ raise FileNotFoundError(f"Reads file 1 not found: {reads1}")
299
+ if not reads2.is_file():
300
+ raise FileNotFoundError(f"Reads file 2 not found: {reads2}")
301
+ if other_ref_fasta and not other_ref_fasta.is_file():
302
+ raise FileNotFoundError(f"Other reference FASTA not found: {other_ref_fasta}")
303
+ if threads <= 0:
304
+ raise ValueError("threads must be a positive integer.")
305
+
306
+ try:
307
+ outdir.mkdir(parents=True, exist_ok=True)
308
+ except Exception as e:
309
+ raise IOError(f"Could not create output directory {outdir}: {e}")
310
+
311
+ # Command construction
312
+ cmd = ["ariba", "run",
313
+ "--assembly_cov", str(assembly_cov),
314
+ "--gene_cov_cutoff", str(gene_cov_cutoff),
315
+ "--gene_nt_extend", str(gene_nt_extend),
316
+ "--local_th", str(local_th),
317
+ "--min_gene_cov", str(min_gene_cov),
318
+ "--assembled_threshold", str(assembled_threshold),
319
+ "--nucmer_min_id", str(nucmer_min_id),
320
+ "--nucmer_min_len", str(nucmer_min_len),
321
+ "--nucmer_breaklen", str(nucmer_breaklen),
322
+ "--nucmer_max_gap", str(nucmer_max_gap),
323
+ "--assembler", assembler,
324
+ "--spades_options", spades_options,
325
+ "--threads", str(threads)]
326
+
327
+ if force: cmd.append("--force")
328
+ if verbose: cmd.append("--verbose")
329
+ if debug: cmd.append("--debug")
330
+ if noclean: cmd.append("--noclean")
331
+ if unique_map_only: cmd.append("--unique_map_only")
332
+ if tmp_dir: cmd.extend(["--tmp_dir", str(tmp_dir)])
333
+ if other_ref_fasta: cmd.extend(["--other_ref_fasta", str(other_ref_fasta)])
334
+ if filestem: cmd.extend(["--filestem", filestem])
335
+
336
+ cmd.extend([str(ref_dir), str(reads1), str(reads2), str(outdir)])
337
+
338
+ # Subprocess execution
339
+ try:
340
+ result = subprocess.run(
341
+ cmd,
342
+ check=True,
343
+ capture_output=True,
344
+ text=True,
345
+ )
346
+ output_files = [str(p) for p in outdir.rglob("*") if p.is_file()]
347
+ return {
348
+ "command_executed": " ".join(cmd),
349
+ "stdout": result.stdout,
350
+ "stderr": result.stderr,
351
+ "output_files": output_files,
352
+ }
353
+ except FileNotFoundError:
354
+ raise RuntimeError("ariba command not found. Please ensure it is in your PATH.")
355
+ except subprocess.CalledProcessError as e:
356
+ return {
357
+ "command_executed": " ".join(cmd),
358
+ "stdout": e.stdout,
359
+ "stderr": e.stderr,
360
+ "error": "ARIBA run failed",
361
+ "return_code": e.returncode,
362
+ }
363
+
364
+ @mcp.tool
365
+ def summary(
366
+ out_prefix: Path,
367
+ in_dirs: List[Path],
368
+ preset: Literal["all", "assembled", "flags", "variants"] = "all",
369
+ cluster_cols: Optional[str] = None,
370
+ col_sep: str = ".",
371
+ row_sep: str = ".",
372
+ force: bool = False,
373
+ verbose: bool = False,
374
+ debug: bool = False,
375
+ ):
376
+ """
377
+ Summarise multiple ARIBA reports into a single file.
378
+
379
+ Args:
380
+ out_prefix: Output file prefix.
381
+ in_dirs: List of directories containing ARIBA results.
382
+ preset: Output preset.
383
+ cluster_cols: Columns to use for clustering.
384
+ col_sep: Separator between column names in output file.
385
+ row_sep: Separator between sample name and cluster name in output file.
386
+ force: Overwrite existing files.
387
+ verbose: Print STDOUT and STDERR from external programs.
388
+ debug: Print debugging messages.
389
+ """
390
+ # Input validation
391
+ if not in_dirs:
392
+ raise ValueError("At least one input directory must be provided in in_dirs.")
393
+ for d in in_dirs:
394
+ if not d.is_dir():
395
+ raise FileNotFoundError(f"Input directory not found: {d}")
396
+ report_path = d / "report.tsv"
397
+ if not report_path.is_file():
398
+ raise FileNotFoundError(f"Expected report file not found in {d}: {report_path}")
399
+
400
+ try:
401
+ out_prefix.parent.mkdir(parents=True, exist_ok=True)
402
+ except Exception as e:
403
+ raise IOError(f"Could not create output directory {out_prefix.parent}: {e}")
404
+
405
+ # Command construction
406
+ cmd = ["ariba", "summary",
407
+ "--preset", preset,
408
+ "--col_sep", col_sep,
409
+ "--row_sep", row_sep]
410
+
411
+ if cluster_cols:
412
+ cmd.extend(["--cluster_cols", cluster_cols])
413
+ if force:
414
+ cmd.append("--force")
415
+ if verbose:
416
+ cmd.append("--verbose")
417
+ if debug:
418
+ cmd.append("--debug")
419
+
420
+ cmd.append(str(out_prefix))
421
+ cmd.extend([str(d) for d in in_dirs])
422
+
423
+ # Subprocess execution
424
+ try:
425
+ result = subprocess.run(
426
+ cmd,
427
+ check=True,
428
+ capture_output=True,
429
+ text=True,
430
+ )
431
+ output_files = [str(p) for p in out_prefix.parent.glob(f"{out_prefix.name}*")]
432
+ return {
433
+ "command_executed": " ".join(cmd),
434
+ "stdout": result.stdout,
435
+ "stderr": result.stderr,
436
+ "output_files": output_files,
437
+ }
438
+ except FileNotFoundError:
439
+ raise RuntimeError("ariba command not found. Please ensure it is in your PATH.")
440
+ except subprocess.CalledProcessError as e:
441
+ return {
442
+ "command_executed": " ".join(cmd),
443
+ "stdout": e.stdout,
444
+ "stderr": e.stderr,
445
+ "error": "ARIBA summary failed",
446
+ "return_code": e.returncode,
447
+ }
448
+
449
+ @mcp.tool
450
+ def version():
451
+ """
452
+ Print the ARIBA version number and exit.
453
+ """
454
+ cmd = ["ariba", "version"]
455
+ try:
456
+ result = subprocess.run(
457
+ cmd,
458
+ check=True,
459
+ capture_output=True,
460
+ text=True,
461
+ )
462
+ return {
463
+ "command_executed": " ".join(cmd),
464
+ "stdout": result.stdout,
465
+ "stderr": result.stderr,
466
+ "version": result.stdout.strip(),
467
+ }
468
+ except FileNotFoundError:
469
+ raise RuntimeError("ariba command not found. Please ensure it is in your PATH.")
470
+ except subprocess.CalledProcessError as e:
471
+ return {
472
+ "command_executed": " ".join(cmd),
473
+ "stdout": e.stdout,
474
+ "stderr": e.stderr,
475
+ "error": "ARIBA version command failed",
476
+ "return_code": e.returncode,
477
+ }
478
+
479
+ @mcp.tool
480
+ def test():
481
+ """
482
+ Run the ARIBA test suite to verify installation and functionality.
483
+ """
484
+ cmd = ["ariba", "test"]
485
+ try:
486
+ # The test command can be long-running and produce a lot of output.
487
+ # We capture it as usual.
488
+ result = subprocess.run(
489
+ cmd,
490
+ check=True,
491
+ capture_output=True,
492
+ text=True,
493
+ )
494
+ return {
495
+ "command_executed": " ".join(cmd),
496
+ "stdout": result.stdout,
497
+ "stderr": result.stderr,
498
+ "test_passed": "OK" in result.stderr,
499
+ }
500
+ except FileNotFoundError:
501
+ raise RuntimeError("ariba command not found. Please ensure it is in your PATH.")
502
+ except subprocess.CalledProcessError as e:
503
+ return {
504
+ "command_executed": " ".join(cmd),
505
+ "stdout": e.stdout,
506
+ "stderr": e.stderr,
507
+ "error": "ARIBA test suite failed",
508
+ "return_code": e.returncode,
509
+ }
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_ariba/app/ariba_shim_server.py ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ from __future__ import annotations
3
+
4
+ import ast
5
+ from pathlib import Path
6
+
7
+ from mcp.server.fastmcp import FastMCP
8
+
9
+
10
+ SOURCE_SERVER = Path('/225040511/project/BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_ariba/app/ariba_server.py')
11
+ SERVER_NAME = 'biosci_ariba'
12
+
13
+
14
+ class _ShimMCP:
15
+ @staticmethod
16
+ def tool():
17
+ def _decorator(fn):
18
+ return fn
19
+ return _decorator
20
+
21
+
22
+ def _load_functions():
23
+ code = SOURCE_SERVER.read_text(encoding="utf-8")
24
+ tree = ast.parse(code, filename=str(SOURCE_SERVER))
25
+ function_names = [n.name for n in tree.body if isinstance(n, ast.FunctionDef) and not n.name.startswith("_")]
26
+ namespace = {
27
+ "__name__": "__mcp_source__",
28
+ "mcp": _ShimMCP(),
29
+ }
30
+ exec(compile(code, str(SOURCE_SERVER), "exec"), namespace, namespace)
31
+ loaded = []
32
+ for name in function_names:
33
+ fn = namespace.get(name)
34
+ if callable(fn):
35
+ loaded.append(fn)
36
+ return loaded
37
+
38
+
39
+ mcp = FastMCP(SERVER_NAME)
40
+ for _fn in _load_functions():
41
+ mcp.tool()(_fn)
42
+
43
+
44
+ if __name__ == "__main__":
45
+ mcp.run(transport="stdio")
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/Dockerfile ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ FROM python:3.10-slim
3
+
4
+ # Install system dependencies
5
+ RUN apt-get update && apt-get install -y default-jre wget curl && apt-get clean && rm -rf /var/lib/apt/lists/*
6
+
7
+ # Install Miniconda
8
+ RUN wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O /tmp/miniconda.sh && bash /tmp/miniconda.sh -b -p /opt/conda && rm /tmp/miniconda.sh
9
+
10
+ # Add conda to PATH
11
+ ENV PATH="/opt/conda/bin:$PATH"
12
+
13
+ # Install arvados-python-client via conda (e.g., from bioconda)
14
+ RUN conda install -c bioconda arvados-python-client -y && conda clean -a
15
+
16
+ # Install Python dependencies
17
+ RUN pip install uv
18
+ RUN uv pip install --system fastmcp
19
+
20
+ # Create app directory
21
+ WORKDIR /app
22
+
23
+ # Copy your MCP server
24
+ COPY app/arvados-python-client_server.py /app/
25
+
26
+ # Create workspace and output directories
27
+ RUN mkdir -p /app/workspace /app/output
28
+
29
+ # Make sure the server script is executable
30
+ RUN chmod +x /app/arvados-python-client_server.py
31
+
32
+ # Expose port for MCP over HTTP (optional)
33
+ EXPOSE 8000
34
+
35
+ # Health check
36
+ HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 CMD python -c "import sys; sys.exit(0)"
37
+
38
+ # Default command runs the MCP server via stdio
39
+ CMD ["python", "/app/arvados-python-client_server.py"]
40
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/app/arvados-python-client_server.py ADDED
@@ -0,0 +1,298 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import subprocess
2
+ import tempfile
3
+ from pathlib import Path
4
+ from typing import List, Optional
5
+
6
+ # MCP-standardized error handling
7
+ class MCPError(Exception):
8
+ def __init__(self, message, stderr):
9
+ self.message = message
10
+ self.stderr = stderr
11
+ super().__init__(self.message)
12
+
13
+ def _run_command(cmd: List[str]):
14
+ """Helper function to run a command and handle exceptions."""
15
+ try:
16
+ command_str = " ".join(cmd)
17
+ result = subprocess.run(
18
+ cmd,
19
+ capture_output=True,
20
+ text=True,
21
+ check=True,
22
+ )
23
+ return {
24
+ "command_executed": command_str,
25
+ "stdout": result.stdout,
26
+ "stderr": result.stderr,
27
+ }
28
+ except FileNotFoundError:
29
+ raise MCPError(f"Command '{cmd[0]}' not found. Is arvados-python-client installed and in the PATH?", "")
30
+ except subprocess.CalledProcessError as e:
31
+ raise MCPError(f"Command '{' '.join(cmd)}' failed with exit code {e.returncode}", e.stderr)
32
+
33
+ @mcp.tool()
34
+ def arv_get(
35
+ object_locator: str,
36
+ destination_path: Optional[Path] = None,
37
+ filename: Optional[str] = None,
38
+ force: bool = False,
39
+ recursive: bool = False,
40
+ progress: bool = False,
41
+ ) -> dict:
42
+ """
43
+ Downloads an object from Arvados Keep to a local path or standard output.
44
+
45
+ Args:
46
+ object_locator: The UUID or portable data hash (PDH) of the object to download.
47
+ destination_path: Local path to save the object. If a directory, the original filename is used. If not provided, content is written to stdout.
48
+ filename: The name of the file to get from a collection.
49
+ force: Overwrite existing files if they exist.
50
+ recursive: Recursively download collections.
51
+ progress: Show a progress bar during download.
52
+
53
+ Returns:
54
+ A dictionary containing the command executed, stdout, stderr, and a list of output files.
55
+ """
56
+ cmd = ["arv-get"]
57
+
58
+ if force:
59
+ cmd.append("--force")
60
+ if recursive:
61
+ cmd.append("--recursive")
62
+ if progress:
63
+ cmd.append("--progress")
64
+ if filename:
65
+ cmd.extend(["--filename", filename])
66
+
67
+ cmd.append(object_locator)
68
+
69
+ output_files = []
70
+ if destination_path:
71
+ # Ensure parent directory exists if a full path is given
72
+ if destination_path.parent and not destination_path.parent.is_dir():
73
+ destination_path.parent.mkdir(parents=True, exist_ok=True)
74
+ cmd.append(str(destination_path))
75
+ output_files.append(str(destination_path))
76
+
77
+ try:
78
+ result = _run_command(cmd)
79
+ result["output_files"] = output_files
80
+ return result
81
+ except MCPError as e:
82
+ return {
83
+ "command_executed": " ".join(cmd),
84
+ "stdout": "",
85
+ "stderr": e.stderr,
86
+ "error": e.message,
87
+ "output_files": [],
88
+ }
89
+
90
+ @mcp.tool()
91
+ def arv_put(
92
+ paths: List[Path],
93
+ name: Optional[str] = None,
94
+ owner: Optional[str] = None,
95
+ portable_data_hash: bool = False,
96
+ storage_class: Optional[str] = None,
97
+ num_retries: int = 5,
98
+ progress: bool = False,
99
+ ) -> dict:
100
+ """
101
+ Uploads local files or directories to Arvados Keep.
102
+
103
+ Args:
104
+ paths: One or more local file or directory paths to upload.
105
+ name: Set the name of the new collection.
106
+ owner: Set the owner project UUID for the new collection.
107
+ portable_data_hash: Use portable data hash to identify the collection.
108
+ storage_class: The storage class for the data (e.g., 'default', 'trash').
109
+ num_retries: Number of times to retry a failed API call.
110
+ progress: Show a progress bar during upload.
111
+
112
+ Returns:
113
+ A dictionary containing the command executed, stdout, and stderr.
114
+ """
115
+ if not paths:
116
+ raise ValueError("At least one path must be provided for upload.")
117
+
118
+ for p in paths:
119
+ if not p.exists():
120
+ raise FileNotFoundError(f"Input path does not exist: {p}")
121
+
122
+ cmd = ["arv-put"]
123
+
124
+ if name:
125
+ cmd.extend(["--name", name])
126
+ if owner:
127
+ cmd.extend(["--owner", owner])
128
+ if portable_data_hash:
129
+ cmd.append("--portable-data-hash")
130
+ if storage_class:
131
+ cmd.extend(["--storage-class", storage_class])
132
+ if progress:
133
+ cmd.append("--progress")
134
+
135
+ # num_retries is a standard arvados client option
136
+ cmd.extend(["--num-retries", str(num_retries)])
137
+
138
+ cmd.extend([str(p) for p in paths])
139
+
140
+ try:
141
+ result = _run_command(cmd)
142
+ result["output_files"] = []
143
+ return result
144
+ except MCPError as e:
145
+ return {
146
+ "command_executed": " ".join(cmd),
147
+ "stdout": "",
148
+ "stderr": e.stderr,
149
+ "error": e.message,
150
+ "output_files": [],
151
+ }
152
+
153
+ @mcp.tool()
154
+ def arv_ls(
155
+ object_locator: str,
156
+ path: Optional[str] = None,
157
+ long_format: bool = False,
158
+ recursive: bool = False,
159
+ human_readable: bool = False,
160
+ ) -> dict:
161
+ """
162
+ Lists the contents of a collection in Arvados Keep.
163
+
164
+ Args:
165
+ object_locator: The UUID or portable data hash (PDH) of the collection to list.
166
+ path: A path within the collection to list.
167
+ long_format: Use a long listing format (similar to ls -l).
168
+ recursive: List subdirectories recursively.
169
+ human_readable: With long_format, print sizes in human-readable format (e.g., 1K, 234M, 2G).
170
+
171
+ Returns:
172
+ A dictionary containing the command executed, stdout (the listing), and stderr.
173
+ """
174
+ cmd = ["arv-ls"]
175
+
176
+ if long_format:
177
+ cmd.append("-l")
178
+ if recursive:
179
+ cmd.append("-R")
180
+ if human_readable:
181
+ cmd.append("-h")
182
+
183
+ cmd.append(object_locator)
184
+ if path:
185
+ cmd.append(path)
186
+
187
+ try:
188
+ result = _run_command(cmd)
189
+ result["output_files"] = []
190
+ return result
191
+ except MCPError as e:
192
+ return {
193
+ "command_executed": " ".join(cmd),
194
+ "stdout": "",
195
+ "stderr": e.stderr,
196
+ "error": e.message,
197
+ "output_files": [],
198
+ }
199
+
200
+ @mcp.tool()
201
+ def arv_copy(
202
+ src_locators: List[str],
203
+ dest_project_uuid: str,
204
+ name: Optional[str] = None,
205
+ num_retries: int = 5,
206
+ ) -> dict:
207
+ """
208
+ Copies one or more objects to a destination project in Arvados.
209
+
210
+ Args:
211
+ src_locators: A list of source object UUIDs or portable data hashes (PDHs).
212
+ dest_project_uuid: The UUID of the destination project.
213
+ name: New name for the object(s). If copying multiple objects, this is ignored.
214
+ num_retries: Number of times to retry a failed API call.
215
+
216
+ Returns:
217
+ A dictionary containing the command executed, stdout, and stderr.
218
+ """
219
+ if not src_locators:
220
+ raise ValueError("At least one source locator must be provided.")
221
+ if not dest_project_uuid:
222
+ raise ValueError("A destination project UUID must be provided.")
223
+
224
+ cmd = ["arv-copy"]
225
+
226
+ if name and len(src_locators) == 1:
227
+ cmd.extend(["--name", name])
228
+
229
+ cmd.extend(["--num-retries", str(num_retries)])
230
+
231
+ cmd.extend(src_locators)
232
+ cmd.append(dest_project_uuid)
233
+
234
+ try:
235
+ result = _run_command(cmd)
236
+ result["output_files"] = []
237
+ return result
238
+ except MCPError as e:
239
+ return {
240
+ "command_executed": " ".join(cmd),
241
+ "stdout": "",
242
+ "stderr": e.stderr,
243
+ "error": e.message,
244
+ "output_files": [],
245
+ }
246
+
247
+ @mcp.tool()
248
+ def arv_tag(
249
+ locator: str,
250
+ tags_to_add: Optional[List[str]] = None,
251
+ keys_to_remove: Optional[List[str]] = None,
252
+ list_tags: bool = False,
253
+ num_retries: int = 5,
254
+ ) -> dict:
255
+ """
256
+ Manipulates tags on an Arvados object. It can list, add, or remove tags.
257
+
258
+ Args:
259
+ locator: The UUID or portable data hash (PDH) of the object to tag.
260
+ tags_to_add: A list of tags to add, in 'key' or 'key=value' format.
261
+ keys_to_remove: A list of tag keys to remove.
262
+ list_tags: If True, lists the existing tags on the object.
263
+ num_retries: Number of times to retry a failed API call.
264
+
265
+ Returns:
266
+ A dictionary containing the command executed, stdout, and stderr.
267
+ """
268
+ cmd = ["arv-tag"]
269
+
270
+ cmd.extend(["--num-retries", str(num_retries)])
271
+
272
+ if list_tags:
273
+ cmd.append("--list")
274
+
275
+ cmd.append(locator)
276
+
277
+ if tags_to_add:
278
+ for tag in tags_to_add:
279
+ if "=" not in tag and " " in tag:
280
+ raise ValueError(f"Invalid tag format: '{tag}'. Use 'key=value' or 'key'.")
281
+ cmd.extend(tags_to_add)
282
+
283
+ if keys_to_remove:
284
+ for key in keys_to_remove:
285
+ cmd.extend(["--remove", key])
286
+
287
+ try:
288
+ result = _run_command(cmd)
289
+ result["output_files"] = []
290
+ return result
291
+ except MCPError as e:
292
+ return {
293
+ "command_executed": " ".join(cmd),
294
+ "stdout": "",
295
+ "stderr": e.stderr,
296
+ "error": e.message,
297
+ "output_files": [],
298
+ }
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/app/arvados-python-client_shim_server.py ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ from __future__ import annotations
3
+
4
+ import ast
5
+ from pathlib import Path
6
+
7
+ from mcp.server.fastmcp import FastMCP
8
+
9
+
10
+ SOURCE_SERVER = Path('/225040511/project/BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/app/arvados-python-client_server.py')
11
+ SERVER_NAME = 'biosci_arvados_python_client'
12
+
13
+
14
+ class _ShimMCP:
15
+ @staticmethod
16
+ def tool():
17
+ def _decorator(fn):
18
+ return fn
19
+ return _decorator
20
+
21
+
22
+ def _load_functions():
23
+ code = SOURCE_SERVER.read_text(encoding="utf-8")
24
+ tree = ast.parse(code, filename=str(SOURCE_SERVER))
25
+ function_names = [n.name for n in tree.body if isinstance(n, ast.FunctionDef) and not n.name.startswith("_")]
26
+ namespace = {
27
+ "__name__": "__mcp_source__",
28
+ "mcp": _ShimMCP(),
29
+ }
30
+ exec(compile(code, str(SOURCE_SERVER), "exec"), namespace, namespace)
31
+ loaded = []
32
+ for name in function_names:
33
+ fn = namespace.get(name)
34
+ if callable(fn):
35
+ loaded.append(fn)
36
+ return loaded
37
+
38
+
39
+ mcp = FastMCP(SERVER_NAME)
40
+ for _fn in _load_functions():
41
+ mcp.tool()(_fn)
42
+
43
+
44
+ if __name__ == "__main__":
45
+ mcp.run(transport="stdio")
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/app/requirements.txt ADDED
@@ -0,0 +1 @@
 
 
1
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/docker-compose.yml ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ version: '3.8'
2
+
3
+ services:
4
+ mcp-arvados-python-client:
5
+ build: .
6
+ image: mcp-arvados-python-client:latest
7
+ container_name: mcp-arvados-python-client
8
+ ports:
9
+ - "8000:8000"
10
+ environment:
11
+ - MCP_SERVER_NAME=arvados-python-client
12
+ volumes:
13
+ - ./workspace:/app/workspace
14
+ - ./output:/app/output
15
+ restart: unless-stopped
16
+ healthcheck:
17
+ test: ["CMD", "python", "-c", "import sys; sys.exit(0)"]
18
+ interval: 30s
19
+ timeout: 10s
20
+ retries: 3
21
+ start_period: 5s
22
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/environment.yaml ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ name: mcp-tool
3
+ channels:
4
+ - bioconda
5
+ - conda-forge
6
+ - defaults
7
+ dependencies:
8
+ - arvados-python-client
9
+ - python=3.10
10
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_arvados-python-client/requirements.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ fastmcp
2
+ mcp
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_augur/requirements.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ fastmcp
2
+ mcp
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/Dockerfile ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ FROM python:3.10-slim
3
+
4
+ # Install system dependencies
5
+ RUN apt-get update && apt-get install -y default-jre wget curl && apt-get clean && rm -rf /var/lib/apt/lists/*
6
+
7
+ # Install Miniconda
8
+ RUN wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O /tmp/miniconda.sh && bash /tmp/miniconda.sh -b -p /opt/conda && rm /tmp/miniconda.sh
9
+
10
+ # Add conda to PATH
11
+ ENV PATH="/opt/conda/bin:$PATH"
12
+
13
+ # Install bcbio-gff via conda (e.g., from bioconda)
14
+ RUN conda install -c bioconda bcbio-gff -y && conda clean -a
15
+
16
+ # Install Python dependencies
17
+ RUN pip install uv
18
+ RUN uv pip install --system fastmcp
19
+
20
+ # Create app directory
21
+ WORKDIR /app
22
+
23
+ # Copy your MCP server
24
+ COPY app/bcbio-gff_server.py /app/
25
+
26
+ # Create workspace and output directories
27
+ RUN mkdir -p /app/workspace /app/output
28
+
29
+ # Make sure the server script is executable
30
+ RUN chmod +x /app/bcbio-gff_server.py
31
+
32
+ # Expose port for MCP over HTTP (optional)
33
+ EXPOSE 8000
34
+
35
+ # Health check
36
+ HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 CMD python -c "import sys; sys.exit(0)"
37
+
38
+ # Default command runs the MCP server via stdio
39
+ CMD ["python", "/app/bcbio-gff_server.py"]
40
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/app/bcbio-gff_server.py ADDED
@@ -0,0 +1,192 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import subprocess
2
+ import tempfile
3
+ from pathlib import Path
4
+ from typing import List, Dict
5
+
6
+ # Note: The 'bcbio-gff' package is primarily a Python library. It also
7
+ # installs several command-line utility scripts. The documentation provided
8
+ # did not include the '--help' output for these scripts. The following MCP
9
+ # tools are based on inferred, common-sense arguments for these utilities.
10
+ # The actual command-line interface may differ.
11
+
12
+ @mcp.tool()
13
+ def gff_to_bed(
14
+ gff_file: Path
15
+ ) -> Dict:
16
+ """
17
+ Converts a GFF file to a BED file using `gff_to_bed.py`.
18
+
19
+ This tool assumes the script takes a single GFF file as input and prints
20
+ the resulting BED formatted data to standard output.
21
+ """
22
+ if not gff_file.is_file():
23
+ raise FileNotFoundError(f"Input GFF file not found: {gff_file}")
24
+
25
+ cmd = ["gff_to_bed.py", str(gff_file)]
26
+
27
+ try:
28
+ result = subprocess.run(
29
+ cmd,
30
+ capture_output=True,
31
+ text=True,
32
+ check=True
33
+ )
34
+
35
+ with tempfile.NamedTemporaryFile(
36
+ mode="w", delete=False, suffix=".bed", prefix="output_"
37
+ ) as tmp_out:
38
+ tmp_out.write(result.stdout)
39
+ output_path = tmp_out.name
40
+
41
+ return {
42
+ "command_executed": " ".join(cmd),
43
+ "stdout": result.stdout,
44
+ "stderr": result.stderr,
45
+ "output_files": [output_path]
46
+ }
47
+ except FileNotFoundError:
48
+ raise RuntimeError("gff_to_bed.py not found. Ensure bcbio-gff scripts are in your PATH.")
49
+ except subprocess.CalledProcessError as e:
50
+ return {
51
+ "command_executed": " ".join(e.cmd),
52
+ "stdout": e.stdout,
53
+ "stderr": e.stderr,
54
+ "output_files": [],
55
+ "error": "Tool execution failed."
56
+ }
57
+
58
+ @mcp.tool()
59
+ def gff_filter_by_attribute(
60
+ gff_file: Path,
61
+ attribute_key: str,
62
+ attribute_value: str
63
+ ) -> Dict:
64
+ """
65
+ Filters a GFF file by a given attribute key-value pair using `gff_filter_by_attribute.py`.
66
+
67
+ This tool assumes the script takes an attribute key, an attribute value, and an
68
+ input GFF file as arguments, and prints the filtered GFF to standard output.
69
+ """
70
+ if not gff_file.is_file():
71
+ raise FileNotFoundError(f"Input GFF file not found: {gff_file}")
72
+
73
+ cmd = ["gff_filter_by_attribute.py", attribute_key, attribute_value, str(gff_file)]
74
+
75
+ try:
76
+ result = subprocess.run(
77
+ cmd,
78
+ capture_output=True,
79
+ text=True,
80
+ check=True
81
+ )
82
+
83
+ with tempfile.NamedTemporaryFile(
84
+ mode="w", delete=False, suffix=".gff", prefix="filtered_"
85
+ ) as tmp_out:
86
+ tmp_out.write(result.stdout)
87
+ output_path = tmp_out.name
88
+
89
+ return {
90
+ "command_executed": " ".join(cmd),
91
+ "stdout": result.stdout,
92
+ "stderr": result.stderr,
93
+ "output_files": [output_path]
94
+ }
95
+ except FileNotFoundError:
96
+ raise RuntimeError("gff_filter_by_attribute.py not found. Ensure bcbio-gff scripts are in your PATH.")
97
+ except subprocess.CalledProcessError as e:
98
+ return {
99
+ "command_executed": " ".join(e.cmd),
100
+ "stdout": e.stdout,
101
+ "stderr": e.stderr,
102
+ "output_files": [],
103
+ "error": "Tool execution failed."
104
+ }
105
+
106
+ @mcp.tool()
107
+ def gff_filter_by_type(
108
+ gff_file: Path,
109
+ feature_type: str
110
+ ) -> Dict:
111
+ """
112
+ Filters a GFF file by a feature type (e.g., 'gene', 'mRNA') using `gff_filter_by_type.py`.
113
+
114
+ This tool assumes the script takes an input GFF file and a feature type as
115
+ arguments, and prints the filtered GFF to standard output.
116
+ """
117
+ if not gff_file.is_file():
118
+ raise FileNotFoundError(f"Input GFF file not found: {gff_file}")
119
+
120
+ cmd = ["gff_filter_by_type.py", str(gff_file), feature_type]
121
+
122
+ try:
123
+ result = subprocess.run(
124
+ cmd,
125
+ capture_output=True,
126
+ text=True,
127
+ check=True
128
+ )
129
+
130
+ with tempfile.NamedTemporaryFile(
131
+ mode="w", delete=False, suffix=".gff", prefix="filtered_type_"
132
+ ) as tmp_out:
133
+ tmp_out.write(result.stdout)
134
+ output_path = tmp_out.name
135
+
136
+ return {
137
+ "command_executed": " ".join(cmd),
138
+ "stdout": result.stdout,
139
+ "stderr": result.stderr,
140
+ "output_files": [output_path]
141
+ }
142
+ except FileNotFoundError:
143
+ raise RuntimeError("gff_filter_by_type.py not found. Ensure bcbio-gff scripts are in your PATH.")
144
+ except subprocess.CalledProcessError as e:
145
+ return {
146
+ "command_executed": " ".join(e.cmd),
147
+ "stdout": e.stdout,
148
+ "stderr": e.stderr,
149
+ "output_files": [],
150
+ "error": "Tool execution failed."
151
+ }
152
+
153
+ @mcp.tool()
154
+ def gff_stats(
155
+ gff_file: Path
156
+ ) -> Dict:
157
+ """
158
+ Calculates statistics on features in a GFF file using `gff_stats.py`.
159
+
160
+ This tool assumes the script takes a single GFF file as input and prints
161
+ statistics to standard output.
162
+ """
163
+ if not gff_file.is_file():
164
+ raise FileNotFoundError(f"Input GFF file not found: {gff_file}")
165
+
166
+ cmd = ["gff_stats.py", str(gff_file)]
167
+
168
+ try:
169
+ result = subprocess.run(
170
+ cmd,
171
+ capture_output=True,
172
+ text=True,
173
+ check=True
174
+ )
175
+
176
+ # The output is statistics text, not a structured file, so we return it directly.
177
+ return {
178
+ "command_executed": " ".join(cmd),
179
+ "stdout": result.stdout,
180
+ "stderr": result.stderr,
181
+ "output_files": []
182
+ }
183
+ except FileNotFoundError:
184
+ raise RuntimeError("gff_stats.py not found. Ensure bcbio-gff scripts are in your PATH.")
185
+ except subprocess.CalledProcessError as e:
186
+ return {
187
+ "command_executed": " ".join(e.cmd),
188
+ "stdout": e.stdout,
189
+ "stderr": e.stderr,
190
+ "output_files": [],
191
+ "error": "Tool execution failed."
192
+ }
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/app/bcbio-gff_shim_server.py ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ from __future__ import annotations
3
+
4
+ import ast
5
+ from pathlib import Path
6
+
7
+ from mcp.server.fastmcp import FastMCP
8
+
9
+
10
+ SOURCE_SERVER = Path('/225040511/project/BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/app/bcbio-gff_server.py')
11
+ SERVER_NAME = 'biosci_bcbio_gff'
12
+
13
+
14
+ class _ShimMCP:
15
+ @staticmethod
16
+ def tool():
17
+ def _decorator(fn):
18
+ return fn
19
+ return _decorator
20
+
21
+
22
+ def _load_functions():
23
+ code = SOURCE_SERVER.read_text(encoding="utf-8")
24
+ tree = ast.parse(code, filename=str(SOURCE_SERVER))
25
+ function_names = [n.name for n in tree.body if isinstance(n, ast.FunctionDef) and not n.name.startswith("_")]
26
+ namespace = {
27
+ "__name__": "__mcp_source__",
28
+ "mcp": _ShimMCP(),
29
+ }
30
+ exec(compile(code, str(SOURCE_SERVER), "exec"), namespace, namespace)
31
+ loaded = []
32
+ for name in function_names:
33
+ fn = namespace.get(name)
34
+ if callable(fn):
35
+ loaded.append(fn)
36
+ return loaded
37
+
38
+
39
+ mcp = FastMCP(SERVER_NAME)
40
+ for _fn in _load_functions():
41
+ mcp.tool()(_fn)
42
+
43
+
44
+ if __name__ == "__main__":
45
+ mcp.run(transport="stdio")
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/app/requirements.txt ADDED
@@ -0,0 +1 @@
 
 
1
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/docker-compose.yml ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ version: '3.8'
2
+
3
+ services:
4
+ mcp-bcbio-gff:
5
+ build: .
6
+ image: mcp-bcbio-gff:latest
7
+ container_name: mcp-bcbio-gff
8
+ ports:
9
+ - "8000:8000"
10
+ environment:
11
+ - MCP_SERVER_NAME=bcbio-gff
12
+ volumes:
13
+ - ./workspace:/app/workspace
14
+ - ./output:/app/output
15
+ restart: unless-stopped
16
+ healthcheck:
17
+ test: ["CMD", "python", "-c", "import sys; sys.exit(0)"]
18
+ interval: 30s
19
+ timeout: 10s
20
+ retries: 3
21
+ start_period: 5s
22
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/environment.yaml ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ name: mcp-tool
3
+ channels:
4
+ - bioconda
5
+ - conda-forge
6
+ - defaults
7
+ dependencies:
8
+ - bcbio-gff
9
+ - python=3.10
10
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bcbio-gff/requirements.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ fastmcp
2
+ mcp
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/Dockerfile ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ FROM python:3.10-slim
3
+
4
+ # Install system dependencies
5
+ RUN apt-get update && apt-get install -y default-jre wget curl && apt-get clean && rm -rf /var/lib/apt/lists/*
6
+
7
+ # Install Miniconda
8
+ RUN wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O /tmp/miniconda.sh && bash /tmp/miniconda.sh -b -p /opt/conda && rm /tmp/miniconda.sh
9
+
10
+ # Add conda to PATH
11
+ ENV PATH="/opt/conda/bin:$PATH"
12
+
13
+ # Install bioblend via conda (e.g., from bioconda)
14
+ RUN conda install -c bioconda bioblend -y && conda clean -a
15
+
16
+ # Install Python dependencies
17
+ RUN pip install uv
18
+ RUN uv pip install --system fastmcp
19
+
20
+ # Create app directory
21
+ WORKDIR /app
22
+
23
+ # Copy your MCP server
24
+ COPY app/bioblend_server.py /app/
25
+
26
+ # Create workspace and output directories
27
+ RUN mkdir -p /app/workspace /app/output
28
+
29
+ # Make sure the server script is executable
30
+ RUN chmod +x /app/bioblend_server.py
31
+
32
+ # Expose port for MCP over HTTP (optional)
33
+ EXPOSE 8000
34
+
35
+ # Health check
36
+ HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 CMD python -c "import sys; sys.exit(0)"
37
+
38
+ # Default command runs the MCP server via stdio
39
+ CMD ["python", "/app/bioblend_server.py"]
40
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/app/__pycache__/bioblend_server.cpython-310.pyc ADDED
Binary file (3.52 kB). View file
 
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/app/bioblend_server.py ADDED
@@ -0,0 +1,111 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import json
2
+ import subprocess
3
+ from pathlib import Path
4
+ from typing import Dict, List, Optional
5
+
6
+ # In a real MCP environment, the 'mcp' object would be imported.
7
+ # For this standalone script, we define a dummy decorator.
8
+ class mcp:
9
+ @staticmethod
10
+ def tool(func):
11
+ return func
12
+
13
+ # Since bioblend is a library, we will need to ensure it is installed
14
+ # in the execution environment. We can simulate a check for it.
15
+ try:
16
+ from bioblend.galaxy import GalaxyInstance
17
+ from bioblend.galaxy.client import Client
18
+ from bioblend.connection import BioBlendConnectionError
19
+ except ImportError:
20
+ raise ImportError("The 'bioblend' library is required to use this tool. Please install it via 'pip install bioblend'.")
21
+
22
+
23
+ @mcp.tool
24
+ def upload_to_galaxy(
25
+ galaxy_url: str,
26
+ api_key: str,
27
+ history_id: str,
28
+ file_path: Path,
29
+ file_type: str = "auto",
30
+ dbkey: str = "?",
31
+ to_posix_lines: bool = True,
32
+ space_to_tab: bool = False,
33
+ ) -> Dict:
34
+ """
35
+ Uploads a dataset to a specified Galaxy history using the bioblend library.
36
+
37
+ This tool acts as a wrapper around the bioblend library's file upload
38
+ functionality, providing a way to programmatically add datasets to a
39
+ Galaxy server from the MCP framework.
40
+
41
+ Args:
42
+ galaxy_url (str): The URL of the Galaxy instance (e.g., 'https://usegalaxy.org').
43
+ api_key (str): Your Galaxy API key for authentication.
44
+ history_id (str): The ID of the history to which the file will be uploaded.
45
+ file_path (Path): The local path to the file to upload.
46
+ file_type (str): The Galaxy datatype for the uploaded file. Defaults to 'auto'.
47
+ dbkey (str): The genome build (dbkey) to associate with the dataset. Defaults to '?'.
48
+ to_posix_lines (bool): If True, convert line endings to POSIX standard. Defaults to True.
49
+ space_to_tab (bool): If True, convert all spaces to tabs in the uploaded file. Defaults to False.
50
+
51
+ Returns:
52
+ Dict: A dictionary containing the execution details and results, including
53
+ the response from the Galaxy API upon successful upload.
54
+ """
55
+ # --- Input Validation ---
56
+ if not galaxy_url:
57
+ raise ValueError("galaxy_url must be provided.")
58
+ if not api_key:
59
+ raise ValueError("api_key must be provided.")
60
+ if not history_id:
61
+ raise ValueError("history_id must be provided.")
62
+
63
+ if not file_path.is_file():
64
+ raise FileNotFoundError(f"Input file not found at: {file_path}")
65
+
66
+ # This tool uses the bioblend library directly, not a command-line subprocess.
67
+ # The 'command_executed' will represent the library call for logging purposes.
68
+ command_executed = (
69
+ f"bioblend.tools.upload_file(file_path='{file_path}', "
70
+ f"history_id='{history_id}', file_type='{file_type}', dbkey='{dbkey}', "
71
+ f"to_posix_lines={to_posix_lines}, space_to_tab={space_to_tab})"
72
+ )
73
+
74
+ stdout_str = ""
75
+ stderr_str = ""
76
+ output_data = {}
77
+
78
+ try:
79
+ # --- Tool Execution ---
80
+ # Instantiate the Galaxy client
81
+ galaxy_instance = GalaxyInstance(url=galaxy_url, key=api_key)
82
+
83
+ # Execute the file upload
84
+ upload_result = galaxy_instance.tools.upload_file(
85
+ str(file_path),
86
+ history_id,
87
+ file_type=file_type,
88
+ dbkey=dbkey,
89
+ to_posix_lines=to_posix_lines,
90
+ space_to_tab=space_to_tab,
91
+ )
92
+
93
+ # Format the successful output
94
+ stdout_str = json.dumps(upload_result, indent=2)
95
+ output_data = upload_result
96
+
97
+ except BioBlendConnectionError as e:
98
+ stderr_str = f"Connection Error: Failed to connect to Galaxy at {galaxy_url}. Details: {e}"
99
+ raise RuntimeError(stderr_str) from e
100
+ except Exception as e:
101
+ # Catch other potential exceptions from the library call
102
+ stderr_str = f"An unexpected error occurred during the bioblend operation: {e}"
103
+ raise RuntimeError(stderr_str) from e
104
+
105
+ return {
106
+ "command_executed": command_executed,
107
+ "stdout": stdout_str,
108
+ "stderr": stderr_str,
109
+ "output_data": output_data,
110
+ "output_files": [] # Output is on the remote Galaxy server, not a local file
111
+ }
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/app/bioblend_shim_server.py ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ from __future__ import annotations
3
+
4
+ import ast
5
+ from pathlib import Path
6
+
7
+ from mcp.server.fastmcp import FastMCP
8
+
9
+
10
+ SOURCE_SERVER = Path('/225040511/project/BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/app/bioblend_server.py')
11
+ SERVER_NAME = 'biosci_bioblend'
12
+
13
+
14
+ class _ShimMCP:
15
+ @staticmethod
16
+ def tool():
17
+ def _decorator(fn):
18
+ return fn
19
+ return _decorator
20
+
21
+
22
+ def _load_functions():
23
+ code = SOURCE_SERVER.read_text(encoding="utf-8")
24
+ tree = ast.parse(code, filename=str(SOURCE_SERVER))
25
+ function_names = [n.name for n in tree.body if isinstance(n, ast.FunctionDef) and not n.name.startswith("_")]
26
+ namespace = {
27
+ "__name__": "__mcp_source__",
28
+ "mcp": _ShimMCP(),
29
+ }
30
+ exec(compile(code, str(SOURCE_SERVER), "exec"), namespace, namespace)
31
+ loaded = []
32
+ for name in function_names:
33
+ fn = namespace.get(name)
34
+ if callable(fn):
35
+ loaded.append(fn)
36
+ return loaded
37
+
38
+
39
+ mcp = FastMCP(SERVER_NAME)
40
+ for _fn in _load_functions():
41
+ mcp.tool()(_fn)
42
+
43
+
44
+ if __name__ == "__main__":
45
+ mcp.run(transport="stdio")
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/app/requirements.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ bioblend==1.9.0
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/docker-compose.yml ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ version: '3.8'
2
+
3
+ services:
4
+ mcp-bioblend:
5
+ build: .
6
+ image: mcp-bioblend:latest
7
+ container_name: mcp-bioblend
8
+ ports:
9
+ - "8000:8000"
10
+ environment:
11
+ - MCP_SERVER_NAME=bioblend
12
+ volumes:
13
+ - ./workspace:/app/workspace
14
+ - ./output:/app/output
15
+ restart: unless-stopped
16
+ healthcheck:
17
+ test: ["CMD", "python", "-c", "import sys; sys.exit(0)"]
18
+ interval: 30s
19
+ timeout: 10s
20
+ retries: 3
21
+ start_period: 5s
22
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/environment.yaml ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ name: mcp-tool
3
+ channels:
4
+ - bioconda
5
+ - conda-forge
6
+ - defaults
7
+ dependencies:
8
+ - bioblend
9
+ - python=3.10
10
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioblend/requirements.txt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ bioblend==1.9.0
2
+ fastmcp
3
+ mcp
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/Dockerfile ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ FROM python:3.10-slim
3
+
4
+ # Install system dependencies
5
+ RUN apt-get update && apt-get install -y default-jre wget curl && apt-get clean && rm -rf /var/lib/apt/lists/*
6
+
7
+ # Install Miniconda
8
+ RUN wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O /tmp/miniconda.sh && bash /tmp/miniconda.sh -b -p /opt/conda && rm /tmp/miniconda.sh
9
+
10
+ # Add conda to PATH
11
+ ENV PATH="/opt/conda/bin:$PATH"
12
+
13
+ # Install bioconductor-affy via conda (e.g., from bioconda)
14
+ RUN conda install -c bioconda bioconductor-affy -y && conda clean -a
15
+
16
+ # Install Python dependencies
17
+ RUN pip install uv
18
+ RUN uv pip install --system fastmcp
19
+
20
+ # Create app directory
21
+ WORKDIR /app
22
+
23
+ # Copy your MCP server
24
+ COPY app/bioconductor-affy_server.py /app/
25
+
26
+ # Create workspace and output directories
27
+ RUN mkdir -p /app/workspace /app/output
28
+
29
+ # Make sure the server script is executable
30
+ RUN chmod +x /app/bioconductor-affy_server.py
31
+
32
+ # Expose port for MCP over HTTP (optional)
33
+ EXPOSE 8000
34
+
35
+ # Health check
36
+ HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 CMD python -c "import sys; sys.exit(0)"
37
+
38
+ # Default command runs the MCP server via stdio
39
+ CMD ["python", "/app/bioconductor-affy_server.py"]
40
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/app/bioconductor-affy_server.py ADDED
@@ -0,0 +1,196 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import subprocess
2
+ import tempfile
3
+ from pathlib import Path
4
+ from typing import Optional, Dict, List
5
+
6
+ # This is a placeholder for the MCP decorator.
7
+ # In a real MCP environment, this would be provided by the MCP framework.
8
+ def tool(func):
9
+ def wrapper(*args, **kwargs):
10
+ # In a real scenario, the MCP framework would handle
11
+ # registration and execution.
12
+ return func(*args, **kwargs)
13
+ return wrapper
14
+ mcp = type("mcp", (), {"tool": tool})
15
+
16
+
17
+ @mcp.tool
18
+ def run_affy_analysis(
19
+ cel_files_dir: Path,
20
+ output_expression_file: Path,
21
+ method: str = "rma",
22
+ pheno_data_file: Optional[Path] = None,
23
+ generate_qc_plot: bool = False,
24
+ qc_plot_file: Optional[Path] = None,
25
+ ) -> Dict:
26
+ """
27
+ Performs exploratory analysis of Affymetrix oligonucleotide arrays.
28
+
29
+ This tool reads raw Affymetrix data from .CEL files, performs background
30
+ correction, normalization, and probe summarization using either the RMA
31
+ (Robust Multi-array Average) or MAS 5.0 algorithm, and outputs a matrix
32
+ of expression values.
33
+
34
+ Args:
35
+ cel_files_dir: Path to the directory containing the input .CEL files.
36
+ output_expression_file: Path to write the output expression matrix (tab-separated).
37
+ method: The processing method to use. Must be one of 'rma' or 'mas5'.
38
+ Defaults to 'rma'.
39
+ pheno_data_file: Optional path to a phenodata file describing the samples.
40
+ This should be a tab-delimited file with row names matching
41
+ the .CEL file names.
42
+ generate_qc_plot: If True, generate a quality control boxplot of the raw
43
+ probe intensities. Defaults to False.
44
+ qc_plot_file: Path to save the QC plot image (e.g., plot.png).
45
+ Required if generate_qc_plot is True.
46
+
47
+ Returns:
48
+ A dictionary containing the executed command, stdout, stderr, and a
49
+ list of output files.
50
+ """
51
+ # 1. Input Validation
52
+ if not cel_files_dir.is_dir():
53
+ raise ValueError(f"Input directory '{cel_files_dir}' does not exist or is not a directory.")
54
+
55
+ if pheno_data_file and not pheno_data_file.is_file():
56
+ raise FileNotFoundError(f"Phenodata file '{pheno_data_file}' not found.")
57
+
58
+ method_lower = method.lower()
59
+ if method_lower not in ["rma", "mas5"]:
60
+ raise ValueError(f"Invalid method '{method}'. Must be 'rma' or 'mas5'.")
61
+
62
+ if generate_qc_plot:
63
+ if qc_plot_file is None:
64
+ raise ValueError("qc_plot_file must be provided when generate_qc_plot is True.")
65
+ if not qc_plot_file.parent.exists():
66
+ qc_plot_file.parent.mkdir(parents=True, exist_ok=True)
67
+ elif qc_plot_file is not None:
68
+ print(f"Warning: qc_plot_file '{qc_plot_file}' is specified but generate_qc_plot is False. The file will not be created.")
69
+
70
+ if not output_expression_file.parent.exists():
71
+ output_expression_file.parent.mkdir(parents=True, exist_ok=True)
72
+
73
+ # 2. R Script Generation
74
+ # This script will be executed by Rscript to perform the analysis.
75
+ r_script_content = f"""
76
+ # Load necessary libraries
77
+ if (!requireNamespace("affy", quietly = TRUE)) {{
78
+ stop("The 'affy' package is not installed. Please install it from Bioconductor.")
79
+ }}
80
+ library(affy)
81
+
82
+ # Parse command-line arguments
83
+ args <- commandArgs(trailingOnly = TRUE)
84
+ if (length(args) < 3) {{
85
+ stop("Usage: Rscript script.R <cel_dir> <output_path> <method> [pheno_path_or_NA] [plot_path_or_NA]", call. = FALSE)
86
+ }}
87
+
88
+ cel_dir <- args[1]
89
+ output_path <- args[2]
90
+ method <- args[3]
91
+ pheno_path <- if (length(args) >= 4 && args[4] != "NA") args[4] else NULL
92
+ plot_path <- if (length(args) >= 5 && args[5] != "NA") args[5] else NULL
93
+
94
+ # --- Main Logic ---
95
+ tryCatch({{
96
+ # Read the data
97
+ cat("Reading CEL files from:", cel_dir, "\\n")
98
+ if (!is.null(pheno_path)) {{
99
+ cat("Using pheno data from:", pheno_path, "\\n")
100
+ # read.AnnotatedDataFrame is deprecated, using read.phenoData is more modern
101
+ pheno <- read.phenoData(pheno_path, header = TRUE, row.names = 1)
102
+ data <- ReadAffy(celfile.path = cel_dir, phenoData = pheno)
103
+ }} else {{
104
+ data <- ReadAffy(celfile.path = cel_dir)
105
+ }}
106
+ cat("Finished reading data.\\n")
107
+
108
+ # Process the data
109
+ cat("Processing data with method:", method, "\\n")
110
+ if (method == "rma") {{
111
+ eset <- rma(data)
112
+ }} else if (method == "mas5") {{
113
+ eset <- mas5(data)
114
+ }} else {{
115
+ stop(paste("Invalid method specified:", method))
116
+ }}
117
+ cat("Finished processing.\\n")
118
+
119
+ # Extract and write expression values
120
+ cat("Writing expression matrix to:", output_path, "\\n")
121
+ expression_matrix <- exprs(eset)
122
+ write.table(expression_matrix, file = output_path, sep = "\\t", quote = FALSE, col.names = NA)
123
+ cat("Done writing expression matrix.\\n")
124
+
125
+ # Generate QC plot if requested
126
+ if (!is.null(plot_path)) {{
127
+ cat("Generating QC plot at:", plot_path, "\\n")
128
+ png(filename = plot_path, width = 800, height = 600)
129
+ boxplot(data, main = "Raw Probe Intensities", las = 2, col="lightblue")
130
+ dev.off()
131
+ cat("Done generating plot.\\n")
132
+ }}
133
+
134
+ cat("Analysis complete.\\n")
135
+ }}, error = function(e) {{
136
+ cat("An error occurred:\\n")
137
+ cat(conditionMessage(e), "\\n")
138
+ quit(status = 1)
139
+ }})
140
+ """
141
+
142
+ # 3. Subprocess Execution
143
+ output_files = {"expression_matrix": str(output_expression_file)}
144
+ cmd = ["Rscript"]
145
+
146
+ try:
147
+ with tempfile.NamedTemporaryFile(mode='w', delete=False, suffix=".R") as r_script_file:
148
+ r_script_file.write(r_script_content)
149
+ r_script_path = r_script_file.name
150
+
151
+ # Prepare arguments for the R script
152
+ pheno_path_arg = str(pheno_data_file.resolve()) if pheno_data_file else "NA"
153
+ plot_path_arg = str(qc_plot_file.resolve()) if generate_qc_plot and qc_plot_file else "NA"
154
+
155
+ cmd.extend([
156
+ r_script_path,
157
+ str(cel_files_dir.resolve()),
158
+ str(output_expression_file.resolve()),
159
+ method_lower,
160
+ pheno_path_arg,
161
+ plot_path_arg
162
+ ])
163
+
164
+ # Execute the command
165
+ process = subprocess.run(
166
+ cmd,
167
+ check=True,
168
+ capture_output=True,
169
+ text=True
170
+ )
171
+
172
+ if generate_qc_plot and qc_plot_file:
173
+ output_files["qc_plot"] = str(qc_plot_file)
174
+
175
+ return {
176
+ "command_executed": " ".join(cmd),
177
+ "stdout": process.stdout,
178
+ "stderr": process.stderr,
179
+ "output_files": output_files
180
+ }
181
+
182
+ except subprocess.CalledProcessError as e:
183
+ return {
184
+ "command_executed": " ".join(cmd),
185
+ "stdout": e.stdout,
186
+ "stderr": e.stderr,
187
+ "error": "R script execution failed.",
188
+ "return_code": e.returncode,
189
+ "output_files": {}
190
+ }
191
+ except Exception as e:
192
+ raise e
193
+ finally:
194
+ # Clean up the temporary R script file
195
+ if 'r_script_path' in locals() and Path(r_script_path).exists():
196
+ Path(r_script_path).unlink()
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/app/bioconductor-affy_shim_server.py ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ from __future__ import annotations
3
+
4
+ import ast
5
+ from pathlib import Path
6
+
7
+ from mcp.server.fastmcp import FastMCP
8
+
9
+
10
+ SOURCE_SERVER = Path('/225040511/project/BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/app/bioconductor-affy_server.py')
11
+ SERVER_NAME = 'biosci_bioconductor_affy'
12
+
13
+
14
+ class _ShimMCP:
15
+ @staticmethod
16
+ def tool():
17
+ def _decorator(fn):
18
+ return fn
19
+ return _decorator
20
+
21
+
22
+ def _load_functions():
23
+ code = SOURCE_SERVER.read_text(encoding="utf-8")
24
+ tree = ast.parse(code, filename=str(SOURCE_SERVER))
25
+ function_names = [n.name for n in tree.body if isinstance(n, ast.FunctionDef) and not n.name.startswith("_")]
26
+ namespace = {
27
+ "__name__": "__mcp_source__",
28
+ "mcp": _ShimMCP(),
29
+ }
30
+ exec(compile(code, str(SOURCE_SERVER), "exec"), namespace, namespace)
31
+ loaded = []
32
+ for name in function_names:
33
+ fn = namespace.get(name)
34
+ if callable(fn):
35
+ loaded.append(fn)
36
+ return loaded
37
+
38
+
39
+ mcp = FastMCP(SERVER_NAME)
40
+ for _fn in _load_functions():
41
+ mcp.tool()(_fn)
42
+
43
+
44
+ if __name__ == "__main__":
45
+ mcp.run(transport="stdio")
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/app/requirements.txt ADDED
@@ -0,0 +1 @@
 
 
1
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/docker-compose.yml ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ version: '3.8'
2
+
3
+ services:
4
+ mcp-bioconductor-affy:
5
+ build: .
6
+ image: mcp-bioconductor-affy:latest
7
+ container_name: mcp-bioconductor-affy
8
+ ports:
9
+ - "8000:8000"
10
+ environment:
11
+ - MCP_SERVER_NAME=bioconductor-affy
12
+ volumes:
13
+ - ./workspace:/app/workspace
14
+ - ./output:/app/output
15
+ restart: unless-stopped
16
+ healthcheck:
17
+ test: ["CMD", "python", "-c", "import sys; sys.exit(0)"]
18
+ interval: 30s
19
+ timeout: 10s
20
+ retries: 3
21
+ start_period: 5s
22
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/environment.yaml ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ name: mcp-tool
3
+ channels:
4
+ - bioconda
5
+ - conda-forge
6
+ - defaults
7
+ dependencies:
8
+ - bioconductor-affy
9
+ - python=3.10
10
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-affy/requirements.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ fastmcp
2
+ mcp
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/Dockerfile ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ FROM python:3.10-slim
3
+
4
+ # Install system dependencies
5
+ RUN apt-get update && apt-get install -y default-jre wget curl && apt-get clean && rm -rf /var/lib/apt/lists/*
6
+
7
+ # Install Miniconda
8
+ RUN wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O /tmp/miniconda.sh && bash /tmp/miniconda.sh -b -p /opt/conda && rm /tmp/miniconda.sh
9
+
10
+ # Add conda to PATH
11
+ ENV PATH="/opt/conda/bin:$PATH"
12
+
13
+ # Install bioconductor-alabaster.sfe via conda (e.g., from bioconda)
14
+ RUN conda install -c bioconda bioconductor-alabaster.sfe -y && conda clean -a
15
+
16
+ # Install Python dependencies
17
+ RUN pip install uv
18
+ RUN uv pip install --system fastmcp
19
+
20
+ # Create app directory
21
+ WORKDIR /app
22
+
23
+ # Copy your MCP server
24
+ COPY app/bioconductor-alabaster.sfe_server.py /app/
25
+
26
+ # Create workspace and output directories
27
+ RUN mkdir -p /app/workspace /app/output
28
+
29
+ # Make sure the server script is executable
30
+ RUN chmod +x /app/bioconductor-alabaster.sfe_server.py
31
+
32
+ # Expose port for MCP over HTTP (optional)
33
+ EXPOSE 8000
34
+
35
+ # Health check
36
+ HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 CMD python -c "import sys; sys.exit(0)"
37
+
38
+ # Default command runs the MCP server via stdio
39
+ CMD ["python", "/app/bioconductor-alabaster.sfe_server.py"]
40
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/app/bioconductor-alabaster.sfe_server.py ADDED
@@ -0,0 +1,213 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import subprocess
2
+ import logging
3
+ from pathlib import Path
4
+ from typing import Dict, Any, Optional
5
+
6
+ # It is assumed that the mcp decorator is available in the execution environment.
7
+ # from mcp import tool
8
+
9
+ # Mock decorator for local testing, not to be included in the final deliverable
10
+ # if 'mcp' not in globals():
11
+ # class mcp:
12
+ # @staticmethod
13
+ # def tool():
14
+ # def decorator(func):
15
+ # return func
16
+ # return decorator
17
+
18
+ logger = logging.getLogger(__name__)
19
+
20
+ @mcp.tool()
21
+ def stage_sfe_object(
22
+ input_rds: Path,
23
+ output_dir: Path,
24
+ path: str = "sfe",
25
+ ) -> Dict[str, Any]:
26
+ """
27
+ Stages a SpatialFeatureExperiment (SFE) object from an RDS file into the alabaster format.
28
+
29
+ This tool wraps the 'stageObject' function from the R package 'alabaster.sfe'.
30
+ It reads an SFE object from a specified RDS file and saves it into a structured
31
+ directory according to the alabaster specification. This serves as the primary
32
+ "save" or "serialize" function for SFE objects in the alabaster framework.
33
+
34
+ Args:
35
+ input_rds: Path to the input RDS file containing the SpatialFeatureExperiment object.
36
+ output_dir: Path to the directory where the staged object assets will be stored.
37
+ path: The relative path within 'output_dir' to save the object. Defaults to "sfe".
38
+
39
+ Returns:
40
+ A dictionary containing the execution details and the path to the main output metadata file.
41
+ """
42
+ # 1. Input Validation
43
+ if not input_rds.is_file():
44
+ raise FileNotFoundError(f"Input RDS file not found at: {input_rds}")
45
+
46
+ if not path:
47
+ raise ValueError("'path' cannot be an empty string.")
48
+
49
+ try:
50
+ output_dir.mkdir(parents=True, exist_ok=True)
51
+ except OSError as e:
52
+ raise OSError(f"Failed to create output directory {output_dir}: {e}")
53
+
54
+ # 2. R script construction
55
+ r_script_content = f"""
56
+ suppressPackageStartupMessages(library(alabaster.sfe))
57
+ suppressPackageStartupMessages(library(SpatialFeatureExperiment))
58
+
59
+ tryCatch({{
60
+ sfe_object <- readRDS('{input_rds.resolve()}')
61
+ info <- alabaster.base::stageObject(sfe_object, dir = '{output_dir.resolve()}', path = '{path}')
62
+ cat(file.path('{output_dir.resolve()}', info$path, "object.json"))
63
+ }}, error = function(e) {{
64
+ message("Error during R script execution: ", e$message)
65
+ quit(status = 1)
66
+ }})
67
+ """
68
+
69
+ # 3. Subprocess Execution
70
+ cmd = ["Rscript", "-e", r_script_content]
71
+ command_executed = " ".join(map(str, cmd))
72
+ logger.info(f"Executing command: {command_executed}")
73
+
74
+ try:
75
+ process = subprocess.run(
76
+ cmd,
77
+ check=True,
78
+ capture_output=True,
79
+ text=True,
80
+ encoding='utf-8'
81
+ )
82
+
83
+ stdout = process.stdout.strip()
84
+ stderr = process.stderr.strip()
85
+
86
+ main_output_file = Path(stdout)
87
+ if not main_output_file.is_file():
88
+ raise FileNotFoundError(
89
+ f"Expected output file not created: {main_output_file}. R stderr: {stderr}"
90
+ )
91
+
92
+ return {
93
+ "command_executed": command_executed,
94
+ "stdout": stdout,
95
+ "stderr": stderr,
96
+ "output_files": {
97
+ "staged_object_json": str(main_output_file)
98
+ }
99
+ }
100
+
101
+ except FileNotFoundError:
102
+ error_msg = "Error: 'Rscript' command not found. Please ensure R is installed and in your system's PATH."
103
+ logger.error(error_msg)
104
+ raise RuntimeError(error_msg) from None
105
+ except subprocess.CalledProcessError as e:
106
+ logger.error(f"Command failed with exit code {e.returncode}")
107
+ logger.error(f"STDOUT: {e.stdout}")
108
+ logger.error(f"STDERR: {e.stderr}")
109
+ raise RuntimeError(
110
+ f"R script execution failed with exit code {e.returncode}.\n"
111
+ f"STDOUT: {e.stdout}\n"
112
+ f"STDERR: {e.stderr}"
113
+ ) from e
114
+
115
+ @mcp.tool()
116
+ def load_sfe_object(
117
+ staged_dir: Path,
118
+ output_rds: Path,
119
+ path: str = "",
120
+ ) -> Dict[str, Any]:
121
+ """
122
+ Loads a SpatialFeatureExperiment (SFE) object from the alabaster format into an RDS file.
123
+
124
+ This tool wraps the 'loadObject' function from the R package 'alabaster.base'.
125
+ It reads an SFE object from a staged directory, reconstructs it in memory, and
126
+ saves the complete object as a single RDS file. This is the "load" or "deserialize"
127
+ counterpart to the staging function.
128
+
129
+ Args:
130
+ staged_dir: Path to the directory containing the staged object assets.
131
+ output_rds: Path for the output RDS file where the loaded SFE object will be saved.
132
+ path: The relative path within 'staged_dir' where the object's 'object.json' is located.
133
+ If empty, it's assumed to be at the root of 'staged_dir'.
134
+
135
+ Returns:
136
+ A dictionary containing the execution details and the path to the output RDS file.
137
+ """
138
+ # 1. Input Validation
139
+ if not staged_dir.is_dir():
140
+ raise NotADirectoryError(f"Staged asset directory not found at: {staged_dir}")
141
+
142
+ metadata_json_path = staged_dir.joinpath(path, "object.json")
143
+ if not metadata_json_path.is_file():
144
+ raise FileNotFoundError(f"Metadata file 'object.json' not found at: {metadata_json_path}")
145
+
146
+ try:
147
+ output_rds.parent.mkdir(parents=True, exist_ok=True)
148
+ except OSError as e:
149
+ raise OSError(f"Failed to create output directory {output_rds.parent}: {e}")
150
+
151
+ # 2. R script construction
152
+ r_script_content = f"""
153
+ suppressPackageStartupMessages(library(alabaster.sfe))
154
+ suppressPackageStartupMessages(library(SpatialFeatureExperiment))
155
+
156
+ tryCatch({{
157
+ sfe_object <- alabaster.base::loadObject(
158
+ '{metadata_json_path.resolve()}',
159
+ project = '{staged_dir.resolve()}'
160
+ )
161
+ saveRDS(sfe_object, file = '{output_rds.resolve()}')
162
+ cat('{output_rds.resolve()}')
163
+ }}, error = function(e) {{
164
+ message("Error during R script execution: ", e$message)
165
+ quit(status = 1)
166
+ }})
167
+ """
168
+
169
+ # 3. Subprocess Execution
170
+ cmd = ["Rscript", "-e", r_script_content]
171
+ command_executed = " ".join(map(str, cmd))
172
+ logger.info(f"Executing command: {command_executed}")
173
+
174
+ try:
175
+ process = subprocess.run(
176
+ cmd,
177
+ check=True,
178
+ capture_output=True,
179
+ text=True,
180
+ encoding='utf-8'
181
+ )
182
+
183
+ stdout = process.stdout.strip()
184
+ stderr = process.stderr.strip()
185
+
186
+ output_file = Path(stdout)
187
+ if not output_file.is_file() or output_file != output_rds.resolve():
188
+ raise FileNotFoundError(
189
+ f"Expected output RDS file not created at {output_rds}. R stderr: {stderr}"
190
+ )
191
+
192
+ return {
193
+ "command_executed": command_executed,
194
+ "stdout": stdout,
195
+ "stderr": stderr,
196
+ "output_files": {
197
+ "output_rds": str(output_file)
198
+ }
199
+ }
200
+
201
+ except FileNotFoundError:
202
+ error_msg = "Error: 'Rscript' command not found. Please ensure R is installed and in your system's PATH."
203
+ logger.error(error_msg)
204
+ raise RuntimeError(error_msg) from None
205
+ except subprocess.CalledProcessError as e:
206
+ logger.error(f"Command failed with exit code {e.returncode}")
207
+ logger.error(f"STDOUT: {e.stdout}")
208
+ logger.error(f"STDERR: {e.stderr}")
209
+ raise RuntimeError(
210
+ f"R script execution failed with exit code {e.returncode}.\n"
211
+ f"STDOUT: {e.stdout}\n"
212
+ f"STDERR: {e.stderr}"
213
+ ) from e
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/app/bioconductor-alabaster.sfe_shim_server.py ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ from __future__ import annotations
3
+
4
+ import ast
5
+ from pathlib import Path
6
+
7
+ from mcp.server.fastmcp import FastMCP
8
+
9
+
10
+ SOURCE_SERVER = Path('/225040511/project/BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/app/bioconductor-alabaster.sfe_server.py')
11
+ SERVER_NAME = 'biosci_bioconductor_alabaster_sfe'
12
+
13
+
14
+ class _ShimMCP:
15
+ @staticmethod
16
+ def tool():
17
+ def _decorator(fn):
18
+ return fn
19
+ return _decorator
20
+
21
+
22
+ def _load_functions():
23
+ code = SOURCE_SERVER.read_text(encoding="utf-8")
24
+ tree = ast.parse(code, filename=str(SOURCE_SERVER))
25
+ function_names = [n.name for n in tree.body if isinstance(n, ast.FunctionDef) and not n.name.startswith("_")]
26
+ namespace = {
27
+ "__name__": "__mcp_source__",
28
+ "mcp": _ShimMCP(),
29
+ }
30
+ exec(compile(code, str(SOURCE_SERVER), "exec"), namespace, namespace)
31
+ loaded = []
32
+ for name in function_names:
33
+ fn = namespace.get(name)
34
+ if callable(fn):
35
+ loaded.append(fn)
36
+ return loaded
37
+
38
+
39
+ mcp = FastMCP(SERVER_NAME)
40
+ for _fn in _load_functions():
41
+ mcp.tool()(_fn)
42
+
43
+
44
+ if __name__ == "__main__":
45
+ mcp.run(transport="stdio")
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/app/requirements.txt ADDED
@@ -0,0 +1 @@
 
 
1
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/docker-compose.yml ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ version: '3.8'
2
+
3
+ services:
4
+ mcp-bioconductor-alabaster.sfe:
5
+ build: .
6
+ image: mcp-bioconductor-alabaster.sfe:latest
7
+ container_name: mcp-bioconductor-alabaster.sfe
8
+ ports:
9
+ - "8000:8000"
10
+ environment:
11
+ - MCP_SERVER_NAME=bioconductor-alabaster.sfe
12
+ volumes:
13
+ - ./workspace:/app/workspace
14
+ - ./output:/app/output
15
+ restart: unless-stopped
16
+ healthcheck:
17
+ test: ["CMD", "python", "-c", "import sys; sys.exit(0)"]
18
+ interval: 30s
19
+ timeout: 10s
20
+ retries: 3
21
+ start_period: 5s
22
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/environment.yaml ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ name: mcp-tool
3
+ channels:
4
+ - bioconda
5
+ - conda-forge
6
+ - defaults
7
+ dependencies:
8
+ - bioconductor-alabaster.sfe
9
+ - python=3.10
10
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-alabaster.sfe/requirements.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ fastmcp
2
+ mcp
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/Dockerfile ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ FROM python:3.10-slim
3
+
4
+ # Install system dependencies
5
+ RUN apt-get update && apt-get install -y default-jre wget curl && apt-get clean && rm -rf /var/lib/apt/lists/*
6
+
7
+ # Install Miniconda
8
+ RUN wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O /tmp/miniconda.sh && bash /tmp/miniconda.sh -b -p /opt/conda && rm /tmp/miniconda.sh
9
+
10
+ # Add conda to PATH
11
+ ENV PATH="/opt/conda/bin:$PATH"
12
+
13
+ # Install bioconductor-biovizbase via conda (e.g., from bioconda)
14
+ RUN conda install -c bioconda bioconductor-biovizbase -y && conda clean -a
15
+
16
+ # Install Python dependencies
17
+ RUN pip install uv
18
+ RUN uv pip install --system fastmcp
19
+
20
+ # Create app directory
21
+ WORKDIR /app
22
+
23
+ # Copy your MCP server
24
+ COPY app/bioconductor-biovizbase_server.py /app/
25
+
26
+ # Create workspace and output directories
27
+ RUN mkdir -p /app/workspace /app/output
28
+
29
+ # Make sure the server script is executable
30
+ RUN chmod +x /app/bioconductor-biovizbase_server.py
31
+
32
+ # Expose port for MCP over HTTP (optional)
33
+ EXPOSE 8000
34
+
35
+ # Health check
36
+ HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 CMD python -c "import sys; sys.exit(0)"
37
+
38
+ # Default command runs the MCP server via stdio
39
+ CMD ["python", "/app/bioconductor-biovizbase_server.py"]
40
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/app/bioconductor-biovizbase_server.py ADDED
@@ -0,0 +1,437 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import subprocess
2
+ from pathlib import Path
3
+ from typing import Optional, List
4
+ import tempfile
5
+ import json
6
+ import re
7
+
8
+ # NOTE: The @mcp.tool decorator is a placeholder for the actual MCP decorator.
9
+ # This code is designed to be used within an MCP environment where `mcp` is defined.
10
+
11
+ class mcp:
12
+ """Placeholder for MCP decorator for standalone validation."""
13
+ @staticmethod
14
+ def tool():
15
+ def decorator(f):
16
+ return f
17
+ return decorator
18
+
19
+ @mcp.tool()
20
+ def biovizbase_get_ideogram(
21
+ genome: str,
22
+ cytoband: bool = True,
23
+ sub_set: Optional[str] = None,
24
+ output_csv: Optional[str] = None
25
+ ) -> dict:
26
+ """
27
+ Fetches chromosome ideogram data for a specified genome build from UCSC.
28
+
29
+ Args:
30
+ genome: Genome build name (e.g., 'hg19', 'mm10'). Must be a valid UCSC genome.
31
+ cytoband: If True, includes cytoband information in the ideogram.
32
+ sub_set: Optional comma-separated list of chromosomes to include (e.g., 'chr1,chr2').
33
+ output_csv: Optional path to save the ideogram data as a CSV file. If not provided, a temporary file will be used.
34
+
35
+ Returns:
36
+ A dictionary containing the command executed, stdout, stderr, and a list of output files.
37
+ """
38
+ if not genome:
39
+ raise ValueError("The 'genome' parameter cannot be empty.")
40
+
41
+ out_path = Path(output_csv) if output_csv else Path(tempfile.mktemp(suffix=".csv"))
42
+
43
+ subset_code = "sub_set = NULL"
44
+ if sub_set:
45
+ # Format the subset string for R: 'chr1,chr2' -> 'c("chr1", "chr2")'
46
+ formatted_sub_set = ','.join([f'"{s.strip()}"' for s in sub_set.split(',')])
47
+ subset_code = f"sub_set = c({formatted_sub_set})"
48
+
49
+ r_script = f"""
50
+ library(biovizBase)
51
+ {subset_code}
52
+ ideogram <- getIdeogram(genome = "{genome}", cytoband = {str(cytoband).upper()}, sub_set = sub_set)
53
+ write.csv(as.data.frame(ideogram), file = "{out_path}", row.names = FALSE)
54
+ """
55
+
56
+ command = ["Rscript", "-e", r_script]
57
+ command_executed = f"Rscript -e '{r_script.strip()}'"
58
+
59
+ try:
60
+ result = subprocess.run(
61
+ command,
62
+ capture_output=True,
63
+ text=True,
64
+ check=True
65
+ )
66
+ return {
67
+ "command_executed": command_executed,
68
+ "stdout": result.stdout,
69
+ "stderr": result.stderr,
70
+ "output_files": [str(out_path)] if out_path.exists() else []
71
+ }
72
+ except subprocess.CalledProcessError as e:
73
+ return {
74
+ "command_executed": command_executed,
75
+ "stdout": e.stdout,
76
+ "stderr": e.stderr,
77
+ "output_files": []
78
+ }
79
+
80
+ @mcp.tool()
81
+ def biovizbase_crunch(
82
+ input_file: str,
83
+ file_format: str,
84
+ which_range: Optional[str] = None,
85
+ output_csv: Optional[str] = None
86
+ ) -> dict:
87
+ """
88
+ Transforms genomic data (e.g., BAM, BED, GTF) into a format suitable for visualization by summarizing ranges.
89
+
90
+ Args:
91
+ input_file: Path to the input genomic data file.
92
+ file_format: Format of the input file (e.g., 'bam', 'bed', 'gtf', 'gff').
93
+ which_range: Genomic range to process, in 'chr:start-end' format (e.g., 'chr1:1000-2000').
94
+ output_csv: Optional path to save the crunched data as a CSV file.
95
+
96
+ Returns:
97
+ A dictionary containing the command executed, stdout, stderr, and a list of output files.
98
+ """
99
+ input_path = Path(input_file)
100
+ if not input_path.exists():
101
+ raise FileNotFoundError(f"Input file not found: {input_file}")
102
+
103
+ if which_range and not re.match(r'^\w+:\d+-\d+$', which_range):
104
+ raise ValueError("The 'which_range' parameter must be in the format 'chr:start-end'.")
105
+
106
+ out_path = Path(output_csv) if output_csv else Path(tempfile.mktemp(suffix=".csv"))
107
+
108
+ range_code = ""
109
+ if which_range:
110
+ chrom, coords = which_range.split(':')
111
+ start, end = coords.split('-')
112
+ range_code = f', which = GRanges("{chrom}", IRanges({start}, {end}))'
113
+
114
+ r_script = f"""
115
+ library(biovizBase)
116
+ library(GenomicRanges)
117
+ library(rtracklayer)
118
+
119
+ if ("{file_format.lower()}" == "bam") {{
120
+ library(Rsamtools)
121
+ data <- readGAlignments("{input_path}")
122
+ }} else {{
123
+ data <- import("{input_path}", format = "{file_format.lower()}")
124
+ }}
125
+
126
+ res <- crunch(data{range_code})
127
+ write.csv(as.data.frame(res), file = "{out_path}", row.names = FALSE)
128
+ """
129
+
130
+ command = ["Rscript", "-e", r_script]
131
+ command_executed = f"Rscript -e '{r_script.strip()}'"
132
+
133
+ try:
134
+ result = subprocess.run(
135
+ command,
136
+ capture_output=True,
137
+ text=True,
138
+ check=True
139
+ )
140
+ return {
141
+ "command_executed": command_executed,
142
+ "stdout": result.stdout,
143
+ "stderr": result.stderr,
144
+ "output_files": [str(out_path)] if out_path.exists() else []
145
+ }
146
+ except subprocess.CalledProcessError as e:
147
+ return {
148
+ "command_executed": command_executed,
149
+ "stdout": e.stdout,
150
+ "stderr": e.stderr,
151
+ "output_files": []
152
+ }
153
+
154
+ @mcp.tool()
155
+ def biovizbase_get_bio_color(
156
+ color_type: str = "DNA"
157
+ ) -> dict:
158
+ """
159
+ Retrieves standard biological color schemes as a JSON object.
160
+
161
+ Args:
162
+ color_type: The category of color scheme. Can be 'DNA', 'RNA', 'AMINO_ACID', 'STRAND', 'CYTOBAND', or a UCSC genome build (e.g., 'hg19').
163
+
164
+ Returns:
165
+ A dictionary containing the command executed, stdout (with JSON color data), and stderr.
166
+ """
167
+ if not color_type:
168
+ raise ValueError("The 'color_type' parameter cannot be empty.")
169
+
170
+ r_script = f"""
171
+ library(biovizBase)
172
+ library(jsonlite)
173
+ cols <- getBioColor(type = "{color_type}")
174
+ cat(jsonlite::toJSON(cols, auto_unbox = TRUE))
175
+ """
176
+
177
+ command = ["Rscript", "-e", r_script]
178
+ command_executed = f"Rscript -e '{r_script.strip()}'"
179
+
180
+ try:
181
+ result = subprocess.run(
182
+ command,
183
+ capture_output=True,
184
+ text=True,
185
+ check=True
186
+ )
187
+ return {
188
+ "command_executed": command_executed,
189
+ "stdout": result.stdout,
190
+ "stderr": result.stderr
191
+ }
192
+ except subprocess.CalledProcessError as e:
193
+ return {
194
+ "command_executed": command_executed,
195
+ "stdout": e.stdout,
196
+ "stderr": e.stderr
197
+ }
198
+
199
+ @mcp.tool()
200
+ def biovizbase_calculate_islands(
201
+ input_file: str,
202
+ output_csv: Optional[str] = None
203
+ ) -> dict:
204
+ """
205
+ Identifies genomic 'islands' (regions of high density/coverage) in alignment or feature data.
206
+
207
+ Args:
208
+ input_file: Path to the input file (e.g., BAM, BED).
209
+ output_csv: Optional path to save the island coordinates as a CSV file.
210
+
211
+ Returns:
212
+ A dictionary containing the command executed, stdout, stderr, and a list of output files.
213
+ """
214
+ input_path = Path(input_file)
215
+ if not input_path.exists():
216
+ raise FileNotFoundError(f"Input file not found: {input_file}")
217
+
218
+ out_path = Path(output_csv) if output_csv else Path(tempfile.mktemp(suffix=".csv"))
219
+
220
+ r_script = f"""
221
+ library(biovizBase)
222
+ library(GenomicRanges)
223
+ library(rtracklayer)
224
+
225
+ if (grepl("\\\\.bam$", "{input_path}", ignore.case = TRUE)) {{
226
+ library(Rsamtools)
227
+ data <- readGAlignments("{input_path}")
228
+ }} else {{
229
+ data <- import("{input_path}")
230
+ }}
231
+
232
+ res <- islands(data)
233
+ write.csv(as.data.frame(res), file = "{out_path}", row.names = FALSE)
234
+ """
235
+
236
+ command = ["Rscript", "-e", r_script]
237
+ command_executed = f"Rscript -e '{r_script.strip()}'"
238
+
239
+ try:
240
+ result = subprocess.run(
241
+ command,
242
+ capture_output=True,
243
+ text=True,
244
+ check=True
245
+ )
246
+ return {
247
+ "command_executed": command_executed,
248
+ "stdout": result.stdout,
249
+ "stderr": result.stderr,
250
+ "output_files": [str(out_path)] if out_path.exists() else []
251
+ }
252
+ except subprocess.CalledProcessError as e:
253
+ return {
254
+ "command_executed": command_executed,
255
+ "stdout": e.stdout,
256
+ "stderr": e.stderr,
257
+ "output_files": []
258
+ }
259
+
260
+ @mcp.tool()
261
+ def biovizbase_add_stepping(
262
+ input_file: str,
263
+ group_name: Optional[str] = None,
264
+ extend: int = 0,
265
+ output_csv: Optional[str] = None
266
+ ) -> dict:
267
+ """
268
+ Calculates 'stepping' levels for genomic ranges to avoid overlaps in visualization.
269
+
270
+ Args:
271
+ input_file: Path to input BED, GFF, or other range-based file.
272
+ group_name: Optional column name in the data to group by for stepping calculation.
273
+ extend: Number of base pairs to extend ranges on each side for overlap calculation. Must be non-negative.
274
+ output_csv: Optional path to save the data with the new 'stepping' column.
275
+
276
+ Returns:
277
+ A dictionary containing the command executed, stdout, stderr, and a list of output files.
278
+ """
279
+ input_path = Path(input_file)
280
+ if not input_path.exists():
281
+ raise FileNotFoundError(f"Input file not found: {input_file}")
282
+ if extend < 0:
283
+ raise ValueError("The 'extend' parameter must be a non-negative integer.")
284
+
285
+ out_path = Path(output_csv) if output_csv else Path(tempfile.mktemp(suffix=".csv"))
286
+ group_code = f', group.name = "{group_name}"' if group_name else ""
287
+
288
+ r_script = f"""
289
+ library(biovizBase)
290
+ library(rtracklayer)
291
+ data <- import("{input_path}")
292
+ res <- addStepping(data{group_code}, extend = {extend})
293
+ write.csv(as.data.frame(res), file = "{out_path}", row.names = FALSE)
294
+ """
295
+
296
+ command = ["Rscript", "-e", r_script]
297
+ command_executed = f"Rscript -e '{r_script.strip()}'"
298
+
299
+ try:
300
+ result = subprocess.run(
301
+ command,
302
+ capture_output=True,
303
+ text=True,
304
+ check=True
305
+ )
306
+ return {
307
+ "command_executed": command_executed,
308
+ "stdout": result.stdout,
309
+ "stderr": result.stderr,
310
+ "output_files": [str(out_path)] if out_path.exists() else []
311
+ }
312
+ except subprocess.CalledProcessError as e:
313
+ return {
314
+ "command_executed": command_executed,
315
+ "stdout": e.stdout,
316
+ "stderr": e.stderr,
317
+ "output_files": []
318
+ }
319
+
320
+ @mcp.tool()
321
+ def biovizbase_rescale_ranges(
322
+ input_file: str,
323
+ to_start: int = 0,
324
+ to_end: int = 1000,
325
+ output_csv: Optional[str] = None
326
+ ) -> dict:
327
+ """
328
+ Rescales genomic coordinates from their original range to a new numeric range.
329
+
330
+ Args:
331
+ input_file: Path to input BED, GFF, or other range-based file.
332
+ to_start: Start of the target numeric range.
333
+ to_end: End of the target numeric range.
334
+ output_csv: Optional path to save the rescaled data.
335
+
336
+ Returns:
337
+ A dictionary containing the command executed, stdout, stderr, and a list of output files.
338
+ """
339
+ input_path = Path(input_file)
340
+ if not input_path.exists():
341
+ raise FileNotFoundError(f"Input file not found: {input_file}")
342
+ if to_start >= to_end:
343
+ raise ValueError("The 'to_start' parameter must be less than 'to_end'.")
344
+
345
+ out_path = Path(output_csv) if output_csv else Path(tempfile.mktemp(suffix=".csv"))
346
+
347
+ r_script = f"""
348
+ library(biovizBase)
349
+ library(rtracklayer)
350
+ data <- import("{input_path}")
351
+ res <- rescale(data, to = c({to_start}, {to_end}))
352
+ write.csv(as.data.frame(res), file = "{out_path}", row.names = FALSE)
353
+ """
354
+
355
+ command = ["Rscript", "-e", r_script]
356
+ command_executed = f"Rscript -e '{r_script.strip()}'"
357
+
358
+ try:
359
+ result = subprocess.run(
360
+ command,
361
+ capture_output=True,
362
+ text=True,
363
+ check=True
364
+ )
365
+ return {
366
+ "command_executed": command_executed,
367
+ "stdout": result.stdout,
368
+ "stderr": result.stderr,
369
+ "output_files": [str(out_path)] if out_path.exists() else []
370
+ }
371
+ except subprocess.CalledProcessError as e:
372
+ return {
373
+ "command_executed": command_executed,
374
+ "stdout": e.stdout,
375
+ "stderr": e.stderr,
376
+ "output_files": []
377
+ }
378
+
379
+ @mcp.tool()
380
+ def biovizbase_shrink_introns(
381
+ input_file: str,
382
+ output_csv: Optional[str] = None
383
+ ) -> dict:
384
+ """
385
+ Shrinks intron regions in transcript data to emphasize exons in visualization.
386
+
387
+ Args:
388
+ input_file: Path to input GTF/GFF file containing transcript models.
389
+ output_csv: Optional path to save the data with shrunk introns.
390
+
391
+ Returns:
392
+ A dictionary containing the command executed, stdout, stderr, and a list of output files.
393
+ """
394
+ input_path = Path(input_file)
395
+ if not input_path.exists():
396
+ raise FileNotFoundError(f"Input file not found: {input_file}")
397
+
398
+ out_path = Path(output_csv) if output_csv else Path(tempfile.mktemp(suffix=".csv"))
399
+
400
+ r_script = f"""
401
+ library(biovizBase)
402
+ library(GenomicFeatures)
403
+ library(rtracklayer)
404
+
405
+ # Create a TxDb object for transcript grouping, which is required for shrink
406
+ txdb <- makeTxDbFromGFF("{input_path}")
407
+ exons <- exonsBy(txdb, by = "tx", use.names = TRUE)
408
+
409
+ res <- shrink(exons)
410
+ # The result is a GRangesList, we flatten it for CSV output
411
+ res_df <- as.data.frame(unlist(res))
412
+ write.csv(res_df, file = "{out_path}", row.names = FALSE)
413
+ """
414
+
415
+ command = ["Rscript", "-e", r_script]
416
+ command_executed = f"Rscript -e '{r_script.strip()}'"
417
+
418
+ try:
419
+ result = subprocess.run(
420
+ command,
421
+ capture_output=True,
422
+ text=True,
423
+ check=True
424
+ )
425
+ return {
426
+ "command_executed": command_executed,
427
+ "stdout": result.stdout,
428
+ "stderr": result.stderr,
429
+ "output_files": [str(out_path)] if out_path.exists() else []
430
+ }
431
+ except subprocess.CalledProcessError as e:
432
+ return {
433
+ "command_executed": command_executed,
434
+ "stdout": e.stdout,
435
+ "stderr": e.stderr,
436
+ "output_files": []
437
+ }
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/app/bioconductor-biovizbase_shim_server.py ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ from __future__ import annotations
3
+
4
+ import ast
5
+ from pathlib import Path
6
+
7
+ from mcp.server.fastmcp import FastMCP
8
+
9
+
10
+ SOURCE_SERVER = Path('/225040511/project/BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/app/bioconductor-biovizbase_server.py')
11
+ SERVER_NAME = 'biosci_bioconductor_biovizbase'
12
+
13
+
14
+ class _ShimMCP:
15
+ @staticmethod
16
+ def tool():
17
+ def _decorator(fn):
18
+ return fn
19
+ return _decorator
20
+
21
+
22
+ def _load_functions():
23
+ code = SOURCE_SERVER.read_text(encoding="utf-8")
24
+ tree = ast.parse(code, filename=str(SOURCE_SERVER))
25
+ function_names = [n.name for n in tree.body if isinstance(n, ast.FunctionDef) and not n.name.startswith("_")]
26
+ namespace = {
27
+ "__name__": "__mcp_source__",
28
+ "mcp": _ShimMCP(),
29
+ }
30
+ exec(compile(code, str(SOURCE_SERVER), "exec"), namespace, namespace)
31
+ loaded = []
32
+ for name in function_names:
33
+ fn = namespace.get(name)
34
+ if callable(fn):
35
+ loaded.append(fn)
36
+ return loaded
37
+
38
+
39
+ mcp = FastMCP(SERVER_NAME)
40
+ for _fn in _load_functions():
41
+ mcp.tool()(_fn)
42
+
43
+
44
+ if __name__ == "__main__":
45
+ mcp.run(transport="stdio")
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/app/requirements.txt ADDED
@@ -0,0 +1 @@
 
 
1
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/docker-compose.yml ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ version: '3.8'
2
+
3
+ services:
4
+ mcp-bioconductor-biovizbase:
5
+ build: .
6
+ image: mcp-bioconductor-biovizbase:latest
7
+ container_name: mcp-bioconductor-biovizbase
8
+ ports:
9
+ - "8000:8000"
10
+ environment:
11
+ - MCP_SERVER_NAME=bioconductor-biovizbase
12
+ volumes:
13
+ - ./workspace:/app/workspace
14
+ - ./output:/app/output
15
+ restart: unless-stopped
16
+ healthcheck:
17
+ test: ["CMD", "python", "-c", "import sys; sys.exit(0)"]
18
+ interval: 30s
19
+ timeout: 10s
20
+ retries: 3
21
+ start_period: 5s
22
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/environment.yaml ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ name: mcp-tool
3
+ channels:
4
+ - bioconda
5
+ - conda-forge
6
+ - defaults
7
+ dependencies:
8
+ - bioconductor-biovizbase
9
+ - python=3.10
10
+
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-biovizbase/requirements.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ fastmcp
2
+ mcp
BioScientist/agent_system/toolbase/mcp_batch_from_manual_txt/mcp_bioconductor-bulksignalr/Dockerfile ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ FROM python:3.10-slim
3
+
4
+ # Install system dependencies
5
+ RUN apt-get update && apt-get install -y default-jre wget curl && apt-get clean && rm -rf /var/lib/apt/lists/*
6
+
7
+ # Install Miniconda
8
+ RUN wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O /tmp/miniconda.sh && bash /tmp/miniconda.sh -b -p /opt/conda && rm /tmp/miniconda.sh
9
+
10
+ # Add conda to PATH
11
+ ENV PATH="/opt/conda/bin:$PATH"
12
+
13
+ # Install bioconductor-bulksignalr via conda (e.g., from bioconda)
14
+ RUN conda install -c bioconda bioconductor-bulksignalr -y && conda clean -a
15
+
16
+ # Install Python dependencies
17
+ RUN pip install uv
18
+ RUN uv pip install --system fastmcp
19
+
20
+ # Create app directory
21
+ WORKDIR /app
22
+
23
+ # Copy your MCP server
24
+ COPY app/bioconductor-bulksignalr_server.py /app/
25
+
26
+ # Create workspace and output directories
27
+ RUN mkdir -p /app/workspace /app/output
28
+
29
+ # Make sure the server script is executable
30
+ RUN chmod +x /app/bioconductor-bulksignalr_server.py
31
+
32
+ # Expose port for MCP over HTTP (optional)
33
+ EXPOSE 8000
34
+
35
+ # Health check
36
+ HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 CMD python -c "import sys; sys.exit(0)"
37
+
38
+ # Default command runs the MCP server via stdio
39
+ CMD ["python", "/app/bioconductor-bulksignalr_server.py"]
40
+