scmcp-shared 0.3.5__tar.gz → 0.3.6__tar.gz

{scmcp_shared-0.3.5 → scmcp_shared-0.3.6}/PKG-INFO

@@ -1,7 +1,10 @@
 Metadata-Version: 2.4
 Name: scmcp_shared
-Version: 0.3.5
+Version: 0.3.6
 Summary: A shared function libray for scmcphub
+Project-URL: Homepage, http://scmcphub.org/
+Project-URL: Repository, https://github.com/scmcphub/scmcp-shared
+Project-URL: Documentation, https://docs.scmcphub.org/
 Author-email: shuang <hsh-me@outlook.com>
 License: BSD 3-Clause License
         
@@ -32,6 +35,7 @@ License: BSD 3-Clause License
         OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
         OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 License-File: LICENSE
+Keywords: AI,agent,bioinformatics,llm,mcp,model context protocol,scRNA-seq,single cell
 Requires-Python: >=3.10
 Requires-Dist: fastmcp>=2.3.4
 Requires-Dist: mcp>=1.8.0

{scmcp_shared-0.3.5 → scmcp_shared-0.3.6}/pyproject.toml

@@ -7,6 +7,18 @@ license = { file = "LICENSE" }
 authors = [
     { name = "shuang", email = "hsh-me@outlook.com" }
 ]
+
+keywords = [
+    "mcp",
+    "model context protocol",
+    "llm",
+    "agent",
+    "scRNA-seq",
+    "single cell",
+    "bioinformatics",
+    "AI"
+]
+
 requires-python = ">=3.10"
 dependencies = [
     "scanpy",
@@ -28,6 +40,12 @@ dev = [
 [tool.hatch.version]
 path = "src/scmcp_shared/__init__.py"  
 
+[project.urls]
+Homepage = "http://scmcphub.org/"
+Repository = "https://github.com/scmcphub/scmcp-shared"
+Documentation = "https://docs.scmcphub.org/"
+
+
 [tool.pytest.ini_options]
 asyncio_mode = "strict"
 asyncio_default_fixture_loop_scope = "function"

scmcp_shared-0.3.6/src/scmcp_shared/__init__.py

@@ -0,0 +1,3 @@
+
+__version__ = "0.3.6"
+

{scmcp_shared-0.3.5 → scmcp_shared-0.3.6}/src/scmcp_shared/schema/io.py

@@ -30,9 +30,9 @@ class ReadModel(BaseModel):
         default=False,
         description="Assume the first column stores row names. This is only necessary if these are not strings: strings in the first column are automatically assumed to be row names."
     )
-    first_column_obs: bool = Field(
-        default=True,
-        description="If True, assume the first column stores observations (cell or barcode) names when provide text file. If False, the data will be transposed."
+    transpose: bool = Field(
+        default=False,
+        description="If True, the data will be transposed."
     )
     backup_url: str = Field(
         default=None,

{scmcp_shared-0.3.5 → scmcp_shared-0.3.6}/src/scmcp_shared/server/io.py

@@ -33,7 +33,7 @@ class ScanpyIOMCP(BaseMCP):
                 elif file.is_file():
                     func_kwargs = filter_args(request, sc.read)
                     adata = sc.read(**func_kwargs)
-                    if not kwargs.get("first_column_obs", True):
+                    if not kwargs.get("transpose", True):
                         adata = adata.T
                 else:
                     raise FileNotFoundError(f"{kwargs['filename']} does not exist")
@@ -47,8 +47,18 @@ class ScanpyIOMCP(BaseMCP):
                 adata.layers["counts"] = adata.X
                 adata.var_names_make_unique()
                 adata.obs_names_make_unique()
+                adata.obs["scmcp_sampleid"] = adinfo.sampleid or ads.active_id
                 ads.set_adata(adata, adinfo=adinfo)
-                return generate_msg(adinfo, adata, ads)
+                return [
+                    {
+                        "sampleid": adinfo.sampleid or ads.active_id, 
+                        "adtype": adinfo.adtype, 
+                        "adata": adata,
+                        "adata.obs_names[:10]": adata.obs_names[:10],
+                        "adata.var_names[:10]": adata.var_names[:10],
+                        "notice": "check obs_names and var_names. transpose the data if needed"
+                    }
+                    ]
             except ToolError as e:
                 raise ToolError(e)
             except Exception as e:

{scmcp_shared-0.3.5 → scmcp_shared-0.3.6}/src/scmcp_shared/server/pp.py

@@ -86,12 +86,12 @@ class ScanpyPreprocessingMCP(BaseMCP):
                     if request.var_max is not None:
                         mask = mask & (adata.var[request.var_key] <= request.var_max)        
                     adata = adata[:, mask]
-                    if request.highly_variable is not None:
-                        adata = adata[:, adata.var.highly_variable]
+                    if request.highly_variable:
+                        adata = adata[:, mask & adata.var.highly_variable]
                     add_op_log(adata, "subset_genes", 
                         {
-                        "var_key": request.var_key, "var_value": request.var_value, 
-                        "var_min": request.var_min, "var_max": request.var_max, "hpv":  request.highly_variable
+                        "var_key": request.var_key,  "var_min": request.var_min, "var_max": request.var_max, 
+                        "hpv":  request.highly_variable
                         }, adinfo
                     )
                 ads.set_adata(adata, adinfo=adinfo)
@@ -133,7 +133,7 @@ class ScanpyPreprocessingMCP(BaseMCP):
         return _calculate_qc_metrics
 
     def _tool_log1p(self):
-        def _log1p(request: Log1PModel, adinfo: self.AdataInfo=self.AdataInfo()):
+        def _log1p(request: Log1PModel=Log1PModel(), adinfo: self.AdataInfo=self.AdataInfo()):
             """Logarithmize the data matrix"""
             try:
                 result = forward_request("pp_log1p", request, adinfo)

{scmcp_shared-0.3.5 → scmcp_shared-0.3.6}/src/scmcp_shared/server/tl.py

@@ -23,7 +23,7 @@ class ScanpyToolsMCP(BaseMCP):
         super().__init__("ScanpyMCP-TL-Server", include_tools, exclude_tools, AdataInfo)
 
     def _tool_tsne(self):
-        def _tsne(request: TSNEModel, adinfo: self.AdataInfo=self.AdataInfo()):
+        def _tsne(request: TSNEModel=TSNEModel(), adinfo: self.AdataInfo=self.AdataInfo()):
             """t-distributed stochastic neighborhood embedding (t-SNE) for visualization"""
             try:
                 result = forward_request("tl_tsne", request, adinfo)
@@ -45,7 +45,7 @@ class ScanpyToolsMCP(BaseMCP):
         return _tsne
 
     def _tool_umap(self):
-        def _umap(request: UMAPModel, adinfo: self.AdataInfo=self.AdataInfo()):
+        def _umap(request: UMAPModel=UMAPModel(), adinfo: self.AdataInfo=self.AdataInfo()):
             """Uniform Manifold Approximation and Projection (UMAP) for visualization"""
             try:
                 result = forward_request("tl_umap", request, adinfo)
@@ -134,7 +134,7 @@ class ScanpyToolsMCP(BaseMCP):
         return _embedding_density
 
     def _tool_leiden(self):
-        def _leiden(request: LeidenModel, adinfo: self.AdataInfo=self.AdataInfo()):
+        def _leiden(request: LeidenModel=LeidenModel(), adinfo: self.AdataInfo=self.AdataInfo()):
             """Leiden clustering algorithm for community detection"""
             try:
                 result = forward_request("tl_leiden", request, adinfo)
@@ -156,7 +156,7 @@ class ScanpyToolsMCP(BaseMCP):
         return _leiden
 
     def _tool_louvain(self):
-        def _louvain(request: LouvainModel, adinfo: self.AdataInfo=self.AdataInfo()):
+        def _louvain(request: LouvainModel=LouvainModel(), adinfo: self.AdataInfo=self.AdataInfo()):
             """Louvain clustering algorithm for community detection"""
             try:
                 result = forward_request("tl_louvain", request, adinfo)

{scmcp_shared-0.3.5 → scmcp_shared-0.3.6}/src/scmcp_shared/server/util.py

@@ -87,7 +87,7 @@ class ScanpyUtilMCP(BaseMCP):
         return _mark_var
 
     def _tool_list_var(self):
-        def _list_var(request: ListVarModel, adinfo: self.AdataInfo=self.AdataInfo()):
+        def _list_var(request: ListVarModel=ListVarModel(), adinfo: self.AdataInfo=self.AdataInfo()):
             """List key columns in adata.var. It should be called for checking when other tools need var key column names as input."""
             try:
                 result = forward_request("ul_list_var", request, adinfo)
@@ -95,7 +95,7 @@ class ScanpyUtilMCP(BaseMCP):
                     return result
                 adata = get_ads().get_adata(adinfo=adinfo)
                 columns = list(adata.var.columns)
-                add_op_log(adata, list_var, {}, adinfo)
+                add_op_log(adata, "list_var", {}, adinfo)
                 return columns
             except ToolError as e:
                 raise ToolError(e)
@@ -115,7 +115,7 @@ class ScanpyUtilMCP(BaseMCP):
                     return result    
                 adata = get_ads().get_adata(adinfo=adinfo)
                 columns = list(adata.obs.columns)
-                add_op_log(adata, list_obs, {}, adinfo)
+                add_op_log(adata, "list_obs", {}, adinfo)
                 return columns
             except ToolError as e:
                 raise ToolError(e)
@@ -135,8 +135,12 @@ class ScanpyUtilMCP(BaseMCP):
                     return result     
                 adata = get_ads().get_adata(adinfo=adinfo)
                 var_names = request.var_names
-                result = {v: v in adata.var_names for v in var_names}
-                add_op_log(adata, check_var, {"var_names": var_names}, adinfo)
+                if adata.raw is not None:
+                    all_var_names = adata.raw.to_adata().var_names
+                else:
+                    all_var_names = adata.var_names
+                result = {v: v in all_var_names for v in var_names}
+                add_op_log(adata, "check_var", {"var_names": var_names}, adinfo)
                 return result
             except ToolError as e:
                 raise ToolError(e)
@@ -157,7 +161,7 @@ class ScanpyUtilMCP(BaseMCP):
                 ads = get_ads()
                 adata = ads.get_adata(adinfo=adinfo)
                 kwargs = {k: v for k, v in request.model_dump().items() if v is not None}
-                merged_adata = adata.concat(list(ads.adata_dic[dtype].values()), **kwargs)
+                merged_adata = adata.concat(ads.adata_dic, **kwargs)
                 ads.adata_dic[dtype] = {}
                 ads.active_id = "merged_adata"
                 add_op_log(merged_adata, ad.concat, kwargs, adinfo)

{scmcp_shared-0.3.5 → scmcp_shared-0.3.6}/src/scmcp_shared/util.py

@@ -104,6 +104,10 @@ def savefig(axes, func=None, **kwargs):
     args = []
     for k,v in kwargs.items():
         if isinstance(v, (tuple, list, set)):
+            v = v[:3]  ## show first 3 elements
+            args.append(f"{k}-{'-'.join([str(i) for i in v])}")
+        elif isinstance(v, dict):
+            v = list(v.keys())[:3]  ## show first 3 elements
             args.append(f"{k}-{'-'.join([str(i) for i in v])}")
         else:
             args.append(f"{k}-{v}")
@@ -209,7 +213,7 @@ def get_ads():
 
 
 def generate_msg(adinfo, adata, ads):
-    return {"sampleid": adinfo.sampleid or ads.active_id, "dtype": adinfo.adtype, "adata": adata}
+    return {"sampleid": adinfo.sampleid or ads.active_id, "adtype": adinfo.adtype, "adata": adata}
 
 
 def sc_like_plot(plot_func, adata, request, adinfo, **kwargs):
@@ -270,3 +274,9 @@ def update_mcp_args(mcp, tool_args : dict):
     tools = mcp._tool_manager._tools.keys()
     for tool in tool_args: 
         _update_args(mcp, tool, tool_args[tool])
+
+
+def check_adata(adata, adinfo, ads):
+    sampleid = adinfo.sampleid or ads.active_id
+    if sampleid != adata.uns["scmcp_sampleid"]:
+        raise ValueError(f"sampleid mismatch: {sampleid} != {adata.uns['scmcp_sampleid']}")

scmcp_shared-0.3.5/src/scmcp_shared/__init__.py

@@ -1,3 +0,0 @@
-
-__version__ = "0.3.5"
-