Merge pull request #33 from virtualcell/add_proper_expression_handling

Extend support for more features of VCell expressions

Author: CodeByDrescher

.github/actions/setup-poetry-env/action.yml

@@ -17,7 +17,7 @@ runs:
 
     - name: Install Poetry
       env:
-        POETRY_VERSION: "1.7.1"
+        POETRY_VERSION: "1.8.5"
       run: curl -sSL https://install.python-poetry.org | python - -y
       shell: bash
 
@@ -32,11 +32,12 @@ runs:
     - name: Load cached venv
       id: cached-poetry-dependencies
       uses: actions/cache@v4
+      if: github.run_attempt == 1
       with:
         path: .venv
         key: venv-${{ runner.os }}-${{ inputs.python-version }}-${{ hashFiles('poetry.lock') }}
 
     - name: Install dependencies
-      if: steps.cached-poetry-dependencies.outputs.cache-hit != 'true'
+      if: github.run_attempt > 1 || steps.cached-poetry-dependencies.outputs.cache-hit != 'true'
       run: poetry install --no-interaction
       shell: bash

.gitignore

@@ -182,3 +182,4 @@ examples/test_output/SimID_*
 examples/notebooks/workspace/
 
 examples/scripts/workspace/
+workspace

examples/notebooks/sbml_workflow.ipynb

@@ -94,7 +94,7 @@
 # Set labels for axes
 ax.set_xlabel("X")
 ax.set_ylabel("Y")
-ax.set_zlabel("Z")  # type: ignore[attr-defined]
+ax.set_zlabel("Z")
 
 # Show the plot
 plt.show()
@@ -138,7 +138,7 @@
 # plot each series on a different plot arranged in a 2x2 grid with a legend from series_legends
 fig, ax = plt.subplots(2, 2, figsize=(10, 10))
 for i, series_array in enumerate(series_arrays):
-    axis = ax[int(i / 2), i % 2]  # type: ignore[index]
+    axis = ax[int(i / 2), i % 2]
     axis.plot(times, series_array[:, 0], label="mean")
     axis.fill_between(times, series_array[:, 1], series_array[:, 2], alpha=0.2)
     axis.set_title(series_legend[i])

examples/scripts/_internal_data_demo.py

@@ -1,12 +1,12 @@
 import matplotlib.pyplot as plt
-from tensorstore._tensorstore import TensorStore  # type: ignore[import-untyped]
+import tensorstore._tensorstore as ts  # type: ignore[import-not-found]
 
 from pyvcell._internal.simdata.n5_data import vcell_n5_datastore
 from pyvcell.sim_results.var_types import NDArray2D
 
 url = "https://vcell-dev.cam.uchc.edu/n5Data/ACowan/4b5ac930c40d5ba.n5"
 dataset_name = "4248805214"
-data_store: TensorStore = vcell_n5_datastore(base_url=url, dataset_name=dataset_name)
+data_store: ts.TensorStore = vcell_n5_datastore(base_url=url, dataset_name=dataset_name)
 print(f"shape = {data_store.shape}")
 
 np_data: NDArray2D = data_store[:, :, 0, 0, 0].read().result()

examples/scripts/_internal_n5_download_demo.py

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "pyvcell"
-version = "0.1.19"
+version = "0.1.20"
 description = "This is the python wrapper for vcell modeling and simulation"
 authors = ["Jim Schaff <schaff@uchc.edu>"]
 repository = "https://github.com/virtualcell/pyvcell"
@@ -32,7 +32,7 @@ matplotlib = "^3.10.0"
 lxml = "^5.3.1"
 imageio = "^2.37.0"
 tensorstore = "^0.1.72"
-libvcell = "^0.0.13"
+libvcell = "^0.0.15"
 trame = "^3.8.1"
 trame-vtk = "^2.8.15"
 trame-vuetify = "^2.8.1"

poetry.lock

@@ -1,7 +1,6 @@
-import tensorstore as ts  # type: ignore[import-untyped]
-from tensorstore._tensorstore import TensorStore  # type: ignore[import-untyped]
+import tensorstore as ts
 
 
-def vcell_n5_datastore(base_url: str, dataset_name: str) -> TensorStore:
+def vcell_n5_datastore(base_url: str, dataset_name: str) -> ts._tensorstore.TensorStore:
     spec = {"driver": "n5", "kvstore": {"driver": "http", "base_url": base_url}, "path": dataset_name}
     return ts.open(spec, read=True).result()

pyproject.toml

@@ -0,0 +1,23 @@
+from libvcell import vcell_infix_to_num_expr_infix
+
+
+def get_numexpr_expression(vcell_expression: str) -> str:
+    # First, get the string from vcell
+    translation_result: bool
+    result_message: str
+    num_expr_expression: str
+    translation_result, result_message, num_expr_expression = vcell_infix_to_num_expr_infix(vcell_expression)
+    if not translation_result:
+        raise ValueError(
+            result_message if result_message is not None else "Unable to convert expression: " + vcell_expression
+        )
+    # NumExpr has a restricted set of what's allowed, in order to protect `eval`
+    sanitized_expression: str = (
+        num_expr_expression.lstrip(" ")
+        .rstrip(" ")
+        .replace(" and ", " & ")
+        .replace(" or ", " | ")
+        .replace(" not ", " ~")
+        .replace("math.", "")
+    )
+    return sanitized_expression

pyvcell/_internal/simdata/n5_data.py

@@ -9,6 +9,7 @@
 import numpy as np
 from numpy._typing import NDArray
 
+from pyvcell._internal.simdata.python_infix import get_numexpr_expression
 from pyvcell.sim_results.var_types import NDArray1D
 
 PYTHON_ENDIANNESS: Literal["little", "big"] = "big"
@@ -350,23 +351,23 @@ def get_data(self, variable: VariableInfo | str, time: float) -> NDArray1D:
 class NamedFunction:
     name: str
     vcell_expression: str
-    python_expression: str
+    num_expr_expression: str
     variables: list[str]
     variable_type: VariableType
 
     def __init__(self, name: str, vcell_expression: str, variable_type: VariableType) -> None:
         self.name = name
         self.vcell_expression = vcell_expression
-        self.python_expression = vcell_expression.replace("^", "**").lstrip(" ").rstrip(" ")
+        self.num_expr_expression = get_numexpr_expression(self.vcell_expression)
         self.variable_type = variable_type
 
         # Parse the python expression into an AST and extract all Name nodes (which represent variables)
-        tree = ast.parse(self.python_expression)
+        tree = ast.parse(self.num_expr_expression)
         self.variables = [node.id for node in ast.walk(tree) if isinstance(node, ast.Name)]
 
     def evaluate(self, variable_bindings: dict[str, NDArray[np.float64]]) -> NDArray[np.float64]:
         ne.set_num_threads(1)
-        expression = self.python_expression
+        expression = self.num_expr_expression
         result = ne.evaluate(expression, local_dict=variable_bindings)
         if not isinstance(result, np.ndarray):
             raise TypeError(f"Expression {expression} did not evaluate to a numpy array")