Vtk #944
Vtk #944
Conversation
📝 WalkthroughWalkthroughThe pull request modifies the Changes
Sequence DiagramsequenceDiagram
participant Input as HDF5 Input File
participant Converter as tovtk.py
participant Output as VTKHDF Output File
Input->>Converter: Read HDF5 File
Converter->>Converter: Determine Dimensionality
alt 2D Case
Converter->>Converter: Process Magnetic Fields (2D)
Converter->>Converter: Process Electric Fields (2D)
else 3D Case
Converter->>Converter: Process Magnetic Fields (3D)
Converter->>Converter: Process Electric Fields (3D)
end
Converter->>Output: Write Converted Data
✨ Finishing Touches
🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
Documentation and Community
|
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
| AMRBoxOffsets = [] | ||
| dataOffsets = [] | ||
|
|
||
| cellData_g = lvl.create_group("CellData") |
Check notice
Code scanning / CodeQL
Unused local variable Note
|
|
||
| cellData_g = lvl.create_group("CellData") | ||
| pointData_g = lvl.create_group("PointData") | ||
| fieldData_g = lvl.create_group("FieldData") |
Check notice
Code scanning / CodeQL
Unused local variable Note
| cellData_g = lvl.create_group("CellData") | ||
| pointData_g = lvl.create_group("PointData") | ||
| fieldData_g = lvl.create_group("FieldData") | ||
| cellDataOffset_g = steps_lvl.create_group("CellDataOffset") |
Check notice
Code scanning / CodeQL
Unused local variable Note
| fieldData_g = lvl.create_group("FieldData") | ||
| cellDataOffset_g = steps_lvl.create_group("CellDataOffset") | ||
| pointDataOffset_g = steps_lvl.create_group("PointDataOffset") | ||
| FieldDataOffset_g = steps_lvl.create_group("FieldDataOffset") |
Check notice
Code scanning / CodeQL
Unused local variable Note
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (4)
pyphare/pyphare/pharesee/tovtk.py (4)
11-43: Add detailed documentation for the Yee grid to primal conversion.The function performs complex averaging operations to convert from Yee grid to primal format. Consider adding:
- Docstring explaining the input parameters and return value
- Mathematical explanation of the averaging operations
- Documentation about the 2D to 3D conversion strategy
Here's a suggested docstring:
def BtoFlatPrimal(ph_bx, ph_by, ph_bz, npx, npy, npz, gn=2): + """Convert magnetic field components from Yee grid to flat primal format. + + Args: + ph_bx, ph_by, ph_bz: Magnetic field components on Yee grid + npx, npy, npz: Number of points in each dimension + gn: Number of ghost nodes (default=2) + + Returns: + numpy.ndarray: Flattened magnetic field in primal format (nbrPoints, 3) + """
45-77: Maintain documentation consistency with BtoFlatPrimal.The function has good inline comments but would benefit from the same level of documentation as suggested for BtoFlatPrimal.
Here's a suggested docstring:
def EtoFlatPrimal(ph_ex, ph_ey, ph_ez, npx, npy, npz, gn=2): + """Convert electric field components from Yee grid to flat primal format. + + Args: + ph_ex, ph_ey, ph_ez: Electric field components on Yee grid + npx, npy, npz: Number of points in each dimension + gn: Number of ghost nodes (default=2) + + Returns: + numpy.ndarray: Flattened electric field in primal format (nbrPoints, 3) + """
79-83: Remove hardcoded 2D assumption.The function hardcodes a 2D case by setting z-coordinates to 0. Consider making it dimension-agnostic for future extensibility.
def boxFromPatch(patch): + """Extract bounding box from patch attributes. + + Args: + patch: HDF5 group containing patch data + + Returns: + list: [x_min, x_max, y_min, y_max, z_min, z_max] + """ lower = patch.attrs["lower"] upper = patch.attrs["upper"] - return [lower[0], upper[0], lower[1], upper[1], 0, 0] # 2D + # Handle both 2D and 3D cases + z_min = lower[2] if len(lower) > 2 else 0 + z_max = upper[2] if len(upper) > 2 else 0 + return [lower[0], upper[0], lower[1], upper[1], z_min, z_max]
185-188: Optimize dictionary key check.Use
not inoperator directly on the dictionary instead of calling.keys().- if phare_lvl_name not in phare_h5["t"][time_str].keys(): + if phare_lvl_name not in phare_h5["t"][time_str]: print(f"no level {ilvl} at time {time}") continue🧰 Tools
🪛 Ruff (0.8.2)
185-185: Use
key not in dictinstead ofkey not in dict.keys()Remove
.keys()(SIM118)
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
pyphare/pyphare/pharesee/tovtk.py(1 hunks)
🧰 Additional context used
🪛 Ruff (0.8.2)
pyphare/pyphare/pharesee/tovtk.py
163-163: Local variable cellData_g is assigned to but never used
Remove assignment to unused variable cellData_g
(F841)
165-165: Local variable fieldData_g is assigned to but never used
Remove assignment to unused variable fieldData_g
(F841)
166-166: Local variable cellDataOffset_g is assigned to but never used
Remove assignment to unused variable cellDataOffset_g
(F841)
168-168: Local variable FieldDataOffset_g is assigned to but never used
Remove assignment to unused variable FieldDataOffset_g
(F841)
185-185: Use key not in dict instead of key not in dict.keys()
Remove .keys()
(SIM118)
⏰ Context from checks skipped due to timeout of 90000ms (6)
- GitHub Check: build (macos-latest)
- GitHub Check: Analyze (cpp)
- GitHub Check: build (macos-14)
- GitHub Check: build (macos-13)
- GitHub Check: Analyze (python)
- GitHub Check: build (ubuntu-latest)
🔇 Additional comments (2)
pyphare/pyphare/pharesee/tovtk.py (2)
1-9: LGTM! Imports are appropriate.The necessary imports for HDF5 handling, numerical operations, and system operations are present.
85-91: Document the reason for adding 1 twice.The function adds 1 twice to each dimension when calculating the number of nodes. This needs explanation.
def nbrNodes(box): + """Calculate number of nodes in each dimension. + + Args: + box: List containing [x_min, x_max, y_min, y_max, z_min, z_max] + + Returns: + tuple: (npx, npy, npz) number of nodes in each dimension + """ lower = box[0], box[2], box[4] upper = box[1], box[3], box[5] + # Adding 1 for converting from cell count to node count + # and another 1 for including the end point npx = upper[0] - lower[0] + 1 + 1 npy = upper[1] - lower[1] + 1 + 1 npz = upper[2] - lower[2] + 1 + 1✅ Verification successful
Documentation accurately explains the double +1 addition
The double addition of 1 is correct and well-documented. The first +1 converts from cell count to node count (as there's always one more node than cells), and the second +1 ensures the endpoint is included, which is required for VTK format compatibility.
🏁 Scripts executed
The following scripts were executed for the analysis:
Error: No shell scripts found
Script:
#!/bin/bash # Search for usages of nbrNodes function rg "nbrNodes" -B 2 -A 2 # Search for similar grid calculations with double +1 pattern rg "([^+]\+\s*1\s*\+\s*1)" -B 2 -A 2 # Look for grid/mesh related files and documentation fd "grid|mesh" --type fLength of output: 3453
Script:
#!/bin/bash # Search for VTK-related code and documentation rg -g "*.py" -g "*.hpp" -g "*.cpp" "vtk|VTK" -B 2 -A 2 # Search for toFlatPrimal implementation rg "toFlatPrimal" -B 5 -A 5 # Search for any documentation about grid/node conversion rg -g "*.py" -g "*.md" "grid.*node|node.*grid" -B 2 -A 2Length of output: 4385
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (5)
pyphare/pyphare/pharesee/tovtk.py (5)
5-8: Add type hints and docstrings to improve code maintainability.Consider adding type hints and docstrings to all functions to improve code maintainability and IDE support. This will help other developers understand the expected input/output types and function purposes.
Example for the
BtoFlatPrimalfunction:-def BtoFlatPrimal(ph_bx, ph_by, ph_bz, npx, npy, npz, gn=2): +from typing import Tuple, ndarray + +def BtoFlatPrimal( + ph_bx: ndarray, + ph_by: ndarray, + ph_bz: ndarray, + npx: int, + npy: int, + npz: int, + gn: int = 2 +) -> ndarray: + """Convert magnetic field components from Yee grid to flat primal format. + + Args: + ph_bx: X-component of magnetic field on Yee grid + ph_by: Y-component of magnetic field on Yee grid + ph_bz: Z-component of magnetic field on Yee grid + npx: Number of points in X direction + npy: Number of points in Y direction + npz: Number of points in Z direction + gn: Number of ghost nodes (default: 2) + + Returns: + ndarray: Flattened magnetic field components in primal format + """Also applies to: 11-11, 45-45, 79-79, 85-85, 94-94
11-43: Consider refactoring field conversion functions to reduce code duplication.The
BtoFlatPrimalandEtoFlatPrimalfunctions share similar structure. Consider extracting common logic into a base function.Example refactor:
def _toFlatPrimal(components: dict, npx: int, npy: int, npz: int, gn: int = 2) -> ndarray: """Base function for converting field components to flat primal format. Args: components: Dictionary containing field components and their averaging rules npx, npy, npz: Number of points in each direction gn: Number of ghost nodes """ nbrPoints = npx * npy * npz result = np.zeros((nbrPoints, 3), dtype="f") # Create pure primal arrays primal = {k: np.zeros((npx, npy, npz), dtype=np.float32) for k in components} # Convert each component using its averaging rule for k, (data, rule) in components.items(): primal[k][:, :, 0] = rule(data, gn) primal[k][:, :, 1] = primal[k][:, :, 0] # Copy to z-dimension # Flatten to output format for i, k in enumerate(components): result[:, i] = primal[k].flatten(order="F") return resultAlso applies to: 45-77
21-32: Add detailed comments explaining the Yee grid to primal conversion process.The ghost node handling and averaging process in the dual direction needs better documentation. Also, document that this implementation assumes 2D data.
Add comments like:
# In Yee grid, B-field components are staggered: # Bx is defined at (i, j+1/2, k+1/2) # By is defined at (i+1/2, j, k+1/2) # Bz is defined at (i+1/2, j+1/2, k) # We average in the dual direction to get values at the primal grid pointsAlso applies to: 55-66
85-91: Document the node number calculation logic.The function adds 1 twice to each dimension. Add comments explaining why this is necessary (e.g., if it's related to cell-centered vs node-centered data).
def nbrNodes(box): + """Calculate number of nodes in each dimension. + + The +1 is added twice because: + 1. Convert from cell count to node count (+1) + 2. [Explain the second +1 here] + """
111-113: Use os.path.join for path construction.Replace string concatenation with
os.path.joinfor more robust path handling across different operating systems.- vtk_fn = f"{data_directory}/{phare_fn}.vtkhdf" + vtk_fn = os.path.join(data_directory, f"{phare_fn}.vtkhdf")
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
pyphare/pyphare/pharesee/tovtk.py(1 hunks)
🧰 Additional context used
🪛 Ruff (0.8.2)
pyphare/pyphare/pharesee/tovtk.py
163-163: Local variable cellData_g is assigned to but never used
Remove assignment to unused variable cellData_g
(F841)
165-165: Local variable fieldData_g is assigned to but never used
Remove assignment to unused variable fieldData_g
(F841)
166-166: Local variable cellDataOffset_g is assigned to but never used
Remove assignment to unused variable cellDataOffset_g
(F841)
168-168: Local variable FieldDataOffset_g is assigned to but never used
Remove assignment to unused variable FieldDataOffset_g
(F841)
185-185: Use key not in dict instead of key not in dict.keys()
Remove .keys()
(SIM118)
223-223: Undefined name b
(F821)
⏰ Context from checks skipped due to timeout of 90000ms (6)
- GitHub Check: build (macos-latest)
- GitHub Check: build (macos-14)
- GitHub Check: Analyze (cpp)
- GitHub Check: build (macos-13)
- GitHub Check: build (ubuntu-latest)
- GitHub Check: Analyze (python)
🔇 Additional comments (2)
pyphare/pyphare/pharesee/tovtk.py (2)
163-168: Remove unused group variables or document their future use.Several group variables are created but never used:
cellData_g,fieldData_g,cellDataOffset_g, andFieldDataOffset_g.🧰 Tools
🪛 Ruff (0.8.2)
163-163: Local variable
cellData_gis assigned to but never usedRemove assignment to unused variable
cellData_g(F841)
165-165: Local variable
fieldData_gis assigned to but never usedRemove assignment to unused variable
fieldData_g(F841)
166-166: Local variable
cellDataOffset_gis assigned to but never usedRemove assignment to unused variable
cellDataOffset_g(F841)
168-168: Local variable
FieldDataOffset_gis assigned to but never usedRemove assignment to unused variable
FieldDataOffset_g(F841)
94-97: Add error handling for command-line arguments.The script assumes the input path is always provided and valid.
works for E, B, n, V 2D
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (4)
pyphare/pyphare/pharesee/tovtk.py (4)
102-106: Consider providing 3D support instead of hardcoding 2D bounds.Currently,
boxFromPatchreturns[lower[0], upper[0], lower[1], upper[1], 0, 0], artificially fixing the z-dimension to 0. If you eventually need 3D support, expand the function to handle[lower[2], upper[2]], or otherwise clarify its 2D-only intent.
155-158: Validate and handle missing command-line arguments.Relying on
sys.argv[1]may cause anIndexErrorif no arguments are passed. Adding a simple argument-count check and verifying file existence can prevent runtime errors.def main(): + if len(sys.argv) < 2: + print("Usage: python tovtk.py <input_h5_file>") + sys.exit(1) path = sys.argv[1] + if not os.path.isfile(path): + print(f"Error: File '{path}' does not exist") + sys.exit(1) phare_h5 = h5py.File(path, "r")
231-231: Use dict membership directly instead of calling.keys().Improve readability and conform to Python best practices by removing redundant
.keys().- if phare_lvl_name not in phare_h5["t"][time_str].keys(): + if phare_lvl_name not in phare_h5["t"][time_str]:🧰 Tools
🪛 Ruff (0.8.2)
231-231: Use
key not in dictinstead ofkey not in dict.keys()Remove
.keys()(SIM118)
297-302: Clarify the dataset shape comment for scalar data.This comment mentions “shape (current_size, 3)” but in the scalar path, the dataset is actually 1D with shape
(current_size,). Updating the comment ensures consistency and avoids confusion.- # dataset already created with shape (current_size,3) + # dataset already created with shape (current_size,) for scalar fields
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
pyphare/pyphare/pharesee/tovtk.py(1 hunks)
🧰 Additional context used
🪛 Ruff (0.8.2)
pyphare/pyphare/pharesee/tovtk.py
209-209: Local variable cellData_g is assigned to but never used
Remove assignment to unused variable cellData_g
(F841)
211-211: Local variable fieldData_g is assigned to but never used
Remove assignment to unused variable fieldData_g
(F841)
212-212: Local variable cellDataOffset_g is assigned to but never used
Remove assignment to unused variable cellDataOffset_g
(F841)
214-214: Local variable FieldDataOffset_g is assigned to but never used
Remove assignment to unused variable FieldDataOffset_g
(F841)
231-231: Use key not in dict instead of key not in dict.keys()
Remove .keys()
(SIM118)
⏰ Context from checks skipped due to timeout of 90000ms (6)
- GitHub Check: Analyze (cpp)
- GitHub Check: Analyze (python)
- GitHub Check: build (macos-14)
- GitHub Check: build (macos-13)
- GitHub Check: build (macos-latest)
- GitHub Check: build (ubuntu-latest)
🔇 Additional comments (1)
pyphare/pyphare/pharesee/tovtk.py (1)
209-214: Remove or document the unused group variables.Variables
cellData_g,fieldData_g,cellDataOffset_g, andFieldDataOffset_gare assigned but never used. They can be removed to reduce clutter or documented if you plan to use them in the future.- cellData_g = lvl.create_group("CellData") pointData_g = lvl.create_group("PointData") - fieldData_g = lvl.create_group("FieldData") - cellDataOffset_g = steps_lvl.create_group("CellDataOffset") pointDataOffset_g = steps_lvl.create_group("PointDataOffset") - FieldDataOffset_g = steps_lvl.create_group("FieldDataOffset")🧰 Tools
🪛 Ruff (0.8.2)
209-209: Local variable
cellData_gis assigned to but never usedRemove assignment to unused variable
cellData_g(F841)
211-211: Local variable
fieldData_gis assigned to but never usedRemove assignment to unused variable
fieldData_g(F841)
212-212: Local variable
cellDataOffset_gis assigned to but never usedRemove assignment to unused variable
cellDataOffset_g(F841)
214-214: Local variable
FieldDataOffset_gis assigned to but never usedRemove assignment to unused variable
FieldDataOffset_g(F841)
There was a problem hiding this comment.
Actionable comments posted: 1
♻️ Duplicate comments (2)
pyphare/pyphare/pharesee/tovtk.py (2)
102-106: 🛠️ Refactor suggestionHardcoded 2D assumption in boxFromPatch needs to be fixed
The function artificially limits functionality by hardcoding z=0, making it incompatible with 3D data.
def boxFromPatch(patch): lower = patch.attrs["lower"] upper = patch.attrs["upper"] - return [lower[0], upper[0], lower[1], upper[1], 0, 0] # 2D + # Handle both 2D and 3D cases + if len(lower) > 2: # 3D case + return [lower[0], upper[0], lower[1], upper[1], lower[2], upper[2]] + else: # 2D case + return [lower[0], upper[0], lower[1], upper[1], 0, 0]
178-178: 🛠️ Refactor suggestionPotential uninitialized variable risk
The
toFlatPrimalvariable might be uninitialized ifprimalFlattenerraises aValueError. This was flagged in previous static analysis.- toFlatPrimal = primalFlattener(phare_fn) + try: + toFlatPrimal = primalFlattener(phare_fn) + except ValueError as e: + print(f"Error: {e}") + sys.exit(1)
🧹 Nitpick comments (5)
pyphare/pyphare/pharesee/tovtk.py (5)
79-100: Vector flattening implementation has redundant flattening operationIn
primalVectorToFlatPrimal, line 95-97 appliesflatten(order="F")to arrays that are already flattened by theprimalScalarToFlatPrimalfunction on lines 91-93.- v[:, 0] = vx.flatten(order="F") - v[:, 1] = vy.flatten(order="F") - v[:, 2] = vz.flatten(order="F") + v[:, 0] = vx + v[:, 1] = vy + v[:, 2] = vz
240-242: Simplify dictionary key checkThe code can be simplified using Python's more idiomatic dictionary membership check.
- if phare_lvl_name not in phare_h5["t"][time_str].keys(): + if phare_lvl_name not in phare_h5["t"][time_str]:🧰 Tools
🪛 Ruff (0.8.2)
240-240: Use
key not in dictinstead ofkey not in dict.keys()Remove
.keys()(SIM118)
275-276: Unnecessary commented codeThere's a commented
# passstatement that should be removed.pointData.resize(current_size + data.shape[0], axis=0) pointData[current_size:, :] = data - # pass
298-298: Unnecessary commented codeAnother commented
# passstatement that should be removed.pointData.resize(current_size + data.shape[0], axis=0) pointData[current_size:] = data - # pass
303-303: Fix comment typoThere's a minor typo in the comment.
- # of of the patch loops at that time + # end of the patch loops at that time
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
pyphare/pyphare/pharesee/tovtk.py(1 hunks)
🧰 Additional context used
🪛 Ruff (0.8.2)
pyphare/pyphare/pharesee/tovtk.py
218-218: Local variable cellData_g is assigned to but never used
Remove assignment to unused variable cellData_g
(F841)
220-220: Local variable fieldData_g is assigned to but never used
Remove assignment to unused variable fieldData_g
(F841)
221-221: Local variable cellDataOffset_g is assigned to but never used
Remove assignment to unused variable cellDataOffset_g
(F841)
223-223: Local variable FieldDataOffset_g is assigned to but never used
Remove assignment to unused variable FieldDataOffset_g
(F841)
240-240: Use key not in dict instead of key not in dict.keys()
Remove .keys()
(SIM118)
⏰ Context from checks skipped due to timeout of 90000ms (6)
- GitHub Check: build (macos-14)
- GitHub Check: build (macos-latest)
- GitHub Check: build (macos-13)
- GitHub Check: Analyze (python)
- GitHub Check: Analyze (cpp)
- GitHub Check: build (ubuntu-latest)
🔇 Additional comments (5)
pyphare/pyphare/pharesee/tovtk.py (5)
45-76: E field transformation uses the correct averaging techniqueThe function properly transforms E field components from Yee grid to primal nodes. The comment on line 65 correctly notes that Ez is already at primal nodes in 2D configuration, so no averaging is needed.
162-165: Improved command-line argument handlingThe added input validation for command-line arguments is a good improvement from the previous version.
218-223: Remove unused group variablesSeveral HDF5 group variables are created but never used:
cellData_g,fieldData_g,cellDataOffset_g, andFieldDataOffset_g. This was flagged in static analysis.- cellData_g = lvl.create_group("CellData") pointData_g = lvl.create_group("PointData") - fieldData_g = lvl.create_group("FieldData") - cellDataOffset_g = steps_lvl.create_group("CellDataOffset") pointDataOffset_g = steps_lvl.create_group("PointDataOffset") - FieldDataOffset_g = steps_lvl.create_group("FieldDataOffset")🧰 Tools
🪛 Ruff (0.8.2)
218-218: Local variable
cellData_gis assigned to but never usedRemove assignment to unused variable
cellData_g(F841)
220-220: Local variable
fieldData_gis assigned to but never usedRemove assignment to unused variable
fieldData_g(F841)
221-221: Local variable
cellDataOffset_gis assigned to but never usedRemove assignment to unused variable
cellDataOffset_g(F841)
223-223: Local variable
FieldDataOffset_gis assigned to but never usedRemove assignment to unused variable
FieldDataOffset_g(F841)
316-317: Script entrypoint looks goodThe main function invocation is properly guarded by the
if __name__ == "__main__"statement, which is a best practice.
11-42:❓ Verification inconclusive
B field transformation logic looks correct but relies on 2D assumptions
The function is well-implemented for converting Yee grid B fields to primal nodes through averaging. However, the code only truly handles 2D data despite the 3D array structures, as seen in:
- The slicing operations only work with 2D data
- Lines 34-36 simply duplicate the same z-slice values
This approach works for the immediate 2D use case but will need refactoring to handle true 3D data in the future.
🏁 Script executed:
#!/bin/bash # Verify if there are any 3D electromagnetic field data in the codebase rg -l "3D" --glob "*.py" | grep -E "field|electromagnetic|B_|E_"Length of output: 64
Action: Confirm 2D Assumptions and Plan for 3D Support
The transformation logic in
pyphare/pyphare/pharesee/tovtk.pyis correctly implemented for converting Yee grid B fields to primal nodes using averaging. However, note that:
- The slicing operations and averaging are effectively handling 2D data.
- Lines 34–36 duplicate the same z-slice values, reinforcing the 2D-only assumption.
- To support true 3D electromagnetic fields in the future, this implementation will require refactoring.
Please manually verify if any parts of the codebase are intended for 3D field processing and consider future changes accordingly.
| return e | ||
|
|
||
|
|
||
| def primalScalarToFlatPrimal(ph_scalar, npx, npy, npz, gn=2): |
Check notice
Code scanning / CodeQL
Explicit returns mixed with implicit (fall through) returns Note
There was a problem hiding this comment.
Actionable comments posted: 4
♻️ Duplicate comments (3)
pyphare/pyphare/pharesee/tovtk.py (3)
155-158:⚠️ Potential issueAdd command-line argument validation
The script assumes the input path is always provided and valid, which could lead to runtime errors.
def main(): + if len(sys.argv) < 2: + print("Usage: python tovtk.py <input_h5_file>") + sys.exit(1) + path = sys.argv[1] + + if not os.path.exists(path): + print(f"Error: File {path} does not exist") + sys.exit(1) + phare_h5 = h5py.File(path, "r")
200-201: 🛠️ Refactor suggestionHardcoded dimensionality in level spacing
The code appends a 0.0 for z-dimension spacing, again showing the 2D-only assumption.
- lvl_spacing = [dl / 2**ilvl for dl in lvl_spacing] + [0.0] + # Calculate spacing based on dimensionality of the data + if len(root_spacing) == 2: + lvl_spacing = [dl / 2**ilvl for dl in lvl_spacing] + [0.0] # Add Z for 2D data + else: + lvl_spacing = [dl / 2**ilvl for dl in lvl_spacing] # Keep as-is for 3D data
102-106: 🛠️ Refactor suggestionHard-coded 2D assumption in boxFromPatch
The function sets z coordinates to 0, assuming 2D data. This prevents proper 3D data handling.
def boxFromPatch(patch): lower = patch.attrs["lower"] upper = patch.attrs["upper"] - return [lower[0], upper[0], lower[1], upper[1], 0, 0] # 2D + # Get dimensionality from the patch + ndim = len(lower) + + if ndim == 2: + return [lower[0], upper[0], lower[1], upper[1], 0, 0] # 2D + elif ndim == 3: + return [lower[0], upper[0], lower[1], upper[1], lower[2], upper[2]] # 3D + else: + raise ValueError(f"Unsupported dimensionality: {ndim}")
🧹 Nitpick comments (6)
pyphare/pyphare/pharesee/tovtk.py (6)
95-97: Redundant flattening operationThe line
v[:, 0] = vx.flatten(order="F")is redundant becausevxis already flattened byprimalScalarToFlatPrimal.- v[:, 0] = vx.flatten(order="F") - v[:, 1] = vy.flatten(order="F") - v[:, 2] = vz.flatten(order="F") + v[:, 0] = vx + v[:, 1] = vy + v[:, 2] = vz
108-114: Add clarifying comment to nbrNodesThe +1+1 in the size calculation is confusing. Add a comment explaining why two +1s are needed.
def nbrNodes(box): lower = box[0], box[2], box[4] upper = box[1], box[3], box[5] - npx = upper[0] - lower[0] + 1 + 1 - npy = upper[1] - lower[1] + 1 + 1 - npz = upper[2] - lower[2] + 1 + 1 + # +1 for inclusive upper bound, +1 for cell to node conversion (n cells = n+1 nodes) + npx = upper[0] - lower[0] + 1 + 1 + npy = upper[1] - lower[1] + 1 + 1 + npz = upper[2] - lower[2] + 1 + 1 return npx, npy, npz
134-142: Optimize max_nbr_levels_in functionThe function can be simplified using max() and a list comprehension for better readability.
def max_nbr_levels_in(phare_h5): - max_nbr_level = 0 times_str = list(phare_h5["t"].keys()) - for time in times_str: - nbrLevels = len(phare_h5["t"][time].keys()) - if max_nbr_level < nbrLevels: - max_nbr_level = nbrLevels - return max_nbr_level + return max([len(phare_h5["t"][time].keys()) for time in times_str], default=0)
257-259: Inconsistent variable namingThe dataset is named
pointData_bbut seems to be used for general data storage.pointData_b = pointData_g.create_dataset( "data", data=data, maxshape=(None, 3) ) # elsewhere - pointData_b.resize(current_size + data.shape[0], axis=0) - pointData_b[current_size:, :] = data + pointData.resize(current_size + data.shape[0], axis=0) + pointData[current_size:, :] = dataAlso applies to: 267-267
263-267: Incorrect comment in dataset resizingThe comment mentions
b.shape[0]but the code usesdata.shape[0].else: # dataset already created with shape (current_size,3) - # we add b.shape[0] points (=npx*npy) to the first dim + # we add data.shape[0] points (=npx*npy) to the first dim # hence need to resize the dataset. pointData_b.resize(current_size + data.shape[0], axis=0) pointData_b[current_size:, :] = data
1-321: Overall structure improvement: extract helper functionsThe main function is quite long and handles many responsibilities. Consider refactoring to improve modularity.
Extract functions for:
- Creating the VTKHDF file structure
- Processing a single patch
- Processing a single time step
This would make the code more maintainable and easier to test.
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (4)
pyphare/pyphare/pharesee/tovtk.py(1 hunks)pyphare/pyphare/pharesee/tovtk.py(5 hunks)pyphare/pyphare/pharesee/tovtk.py(3 hunks)pyphare/pyphare/pharesee/tovtk.py(5 hunks)
🚧 Files skipped from review as they are similar to previous changes (2)
- pyphare/pyphare/pharesee/tovtk.py
- pyphare/pyphare/pharesee/tovtk.py
⏰ Context from checks skipped due to timeout of 90000ms (7)
- GitHub Check: build (macos-latest)
- GitHub Check: build (macos-14)
- GitHub Check: Analyze (cpp)
- GitHub Check: build (ubuntu-latest, clang)
- GitHub Check: build (ubuntu-latest, gcc)
- GitHub Check: Analyze (python)
- GitHub Check: build (macos-13)
🔇 Additional comments (1)
pyphare/pyphare/pharesee/tovtk.py (1)
209-214: Remove unused group variablesSeveral group variables are created but never used:
cellData_g,fieldData_g,cellDataOffset_g, andFieldDataOffset_g.- cellData_g = lvl.create_group("CellData") pointData_g = lvl.create_group("PointData") - fieldData_g = lvl.create_group("FieldData") - cellDataOffset_g = steps_lvl.create_group("CellDataOffset") pointDataOffset_g = steps_lvl.create_group("PointDataOffset") - FieldDataOffset_g = steps_lvl.create_group("FieldDataOffset")
| by[:, :, 0] = ( | ||
| ph_by[gn - 1 : -gn + 1, gn:-gn][1:, :] | ||
| + ph_by[gn - 1 : -gn + 1, gn:-gn][:-1, :] | ||
| ) * 0.5 | ||
|
|
||
| bz[:, :, 0] = 0.5 * ( | ||
| 0.5 * (ph_bz[domainP1, domain][1:, :] + ph_bz[domainP1, domain][:-1, :]) | ||
| + 0.5 * (ph_bz[domain, domainP1][:, 1:] + ph_bz[domain, domainP1][:, :-1]) | ||
| ) | ||
|
|
||
| bx[:, :, 1] = bx[:, :, 0] | ||
| by[:, :, 1] = by[:, :, 0] | ||
| bz[:, :, 1] = bz[:, :, 0] | ||
|
|
||
| elif ndim_from(npx, npy, npz) == 3: | ||
| # Bx is (primal, dual, dual) | ||
| print("bx", ph_bx.shape) | ||
| print( | ||
| "bx[domain, domainP1, domain][:,1:,:]", | ||
| ph_bx[domain, domainP1, domain][:, 1:, :].shape, | ||
| ) | ||
| print( | ||
| "bx[domain, domainP1, domain][:,:-1,:]", | ||
| ph_bx[domain, domainP1, domain][:, :-1, :].shape, | ||
| ) | ||
| bx[:, :, :] = 0.5 * ( | ||
| 0.5 | ||
| * ( | ||
| ph_bx[domain, domainP1, domain][:, 1:, :] | ||
| + ph_bx[domain, domainP1, domain][:, :-1, :] | ||
| ) | ||
| + 0.5 | ||
| * ( |
There was a problem hiding this comment.
🛠️ Refactor suggestion
Similar dimensionality issue in EtoFlatPrimal
Like BtoFlatPrimal, this function also handles primarily 2D data. Additionally, the code comment on line 65 explicitly states "ez already primal in 2D".
Apply a similar pattern as suggested for BtoFlatPrimal to handle both 2D and 3D cases appropriately.
| + ph_bx[domain, domain, domainP1][:, :, :-1] | ||
| ) | ||
| ) | ||
| # By is (dual, primal, dual) | ||
| by[:, :, :] = 0.5 * ( | ||
| 0.5 | ||
| * ( |
There was a problem hiding this comment.
🛠️ Refactor suggestion
primalScalarToFlatPrimal uses hardcoded 2D handling
The function duplicates the 2D array for z-dimension without checking if the data is actually 3D.
def primalScalarToFlatPrimal(ph_scalar, npx, npy, npz, gn=2):
-
+ ndim = 2 if npz <= 2 else 3
scalar3d = np.zeros((npx, npy, npz), dtype="f")
- scalar3d[:, :, 0] = ph_scalar[gn:-gn, gn:-gn]
- scalar3d[:, :, 1] = ph_scalar[gn:-gn, gn:-gn]
+ if ndim == 2:
+ scalar3d[:, :, 0] = ph_scalar[gn:-gn, gn:-gn]
+ # Duplicate the 2D slice for all z-slices
+ for i in range(1, npz):
+ scalar3d[:, :, i] = scalar3d[:, :, 0]
+ else:
+ # Handle 3D case when implemented
+ raise NotImplementedError("3D support not yet implemented")
return scalar3d.flatten(order="F")Committable suggestion skipped: line range outside the PR's diff.
| def ndim_from(npx, npy, npz): | ||
| if npx > 2 and npy > 2 and npz > 2: | ||
| return 3 | ||
| elif npx > 2 and npy > 2 and npz == 2: | ||
| return 2 | ||
| elif npx > 2 and npy == 2 and npz == 2: | ||
| return 1 | ||
| else: | ||
| raise ValueError( | ||
| f" cannot infer dimension from (npx, npy, npz) = {npx} {npy} {npz}" | ||
| ) | ||
|
|
||
|
|
||
| def BtoFlatPrimal(ph_bx, ph_by, ph_bz, npx, npy, npz, gn=2): | ||
|
|
||
| nbrPoints = npx * npy * npz | ||
| b = np.zeros((nbrPoints, 3), dtype="f") | ||
|
|
||
| # pure primal arrays | ||
| bx = np.zeros((npx, npy, npz), dtype=np.float32) | ||
| by = np.zeros((npx, npy, npz), dtype=np.float32) | ||
| bz = np.zeros((npx, npy, npz), dtype=np.float32) | ||
|
|
||
| domainP1 = slice(gn - 1, -gn + 1) | ||
| domain = slice(gn, -gn) | ||
|
|
||
| # converts yee to pure primal | ||
| # we average in the dual direction so we need one extract ghost data in that direction | ||
|
|
||
| if ndim_from(npx, npy, npz) == 2: | ||
| bx[:, :, 0] = ( | ||
| ph_bx[gn:-gn, gn - 1 : -gn + 1][:, 1:] | ||
| + ph_bx[gn:-gn, gn - 1 : -gn + 1][:, :-1] |
There was a problem hiding this comment.
🛠️ Refactor suggestion
Dimensionality handling in BtoFlatPrimal needs improvement
The function seems to handle primarily 2D data (z-dimension) by duplicating 2D slices, but the API accepts npz parameter and creates 3D arrays. Consider implementing proper 3D support for full-dimensional data handling.
def BtoFlatPrimal(ph_bx, ph_by, ph_bz, npx, npy, npz, gn=2):
+ ndim = 2 if npz <= 2 else 3
nbrPoints = npx * npy * npz
b = np.zeros((nbrPoints, 3), dtype="f")
# pure primal arrays
bx = np.zeros((npx, npy, npz), dtype=np.float32)
by = np.zeros((npx, npy, npz), dtype=np.float32)
bz = np.zeros((npx, npy, npz), dtype=np.float32)
# converts yee to pure primal
# we average in the dual direction so we need one extract ghost data in that direction
- bx[:, :, 0] = (
- ph_bx[gn:-gn, gn - 1 : -gn + 1][:, 1:] + ph_bx[gn:-gn, gn - 1 : -gn + 1][:, :-1]
- ) * 0.5
- by[:, :, 0] = (
- ph_by[gn - 1 : -gn + 1, gn:-gn][1:, :] + ph_by[gn - 1 : -gn + 1, gn:-gn][:-1, :]
- ) * 0.5
- bz[:, :, 0] = (
- ph_bz[gn - 1 : -gn + 1, gn - 1 : -gn + 1][1:, 1:]
- + ph_bz[gn - 1 : -gn + 1, gn - 1 : -gn + 1][:-1, :-1]
- ) * 0.5
-
- bx[:, :, 1] = bx[:, :, 0]
- by[:, :, 1] = by[:, :, 0]
- bz[:, :, 1] = bz[:, :, 0]
+ if ndim == 2:
+ bx[:, :, 0] = (
+ ph_bx[gn:-gn, gn - 1 : -gn + 1][:, 1:] + ph_bx[gn:-gn, gn - 1 : -gn + 1][:, :-1]
+ ) * 0.5
+ by[:, :, 0] = (
+ ph_by[gn - 1 : -gn + 1, gn:-gn][1:, :] + ph_by[gn - 1 : -gn + 1, gn:-gn][:-1, :]
+ ) * 0.5
+ bz[:, :, 0] = (
+ ph_bz[gn - 1 : -gn + 1, gn - 1 : -gn + 1][1:, 1:]
+ + ph_bz[gn - 1 : -gn + 1, gn - 1 : -gn + 1][:-1, :-1]
+ ) * 0.5
+
+ # Duplicate 2D data for pseudo-3D representation
+ for i in range(1, npz):
+ bx[:, :, i] = bx[:, :, 0]
+ by[:, :, i] = by[:, :, 0]
+ bz[:, :, i] = bz[:, :, 0]
+ else:
+ # Add 3D handling code here when needed
+ raise NotImplementedError("3D support not yet implemented")📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| def ndim_from(npx, npy, npz): | |
| if npx > 2 and npy > 2 and npz > 2: | |
| return 3 | |
| elif npx > 2 and npy > 2 and npz == 2: | |
| return 2 | |
| elif npx > 2 and npy == 2 and npz == 2: | |
| return 1 | |
| else: | |
| raise ValueError( | |
| f" cannot infer dimension from (npx, npy, npz) = {npx} {npy} {npz}" | |
| ) | |
| def BtoFlatPrimal(ph_bx, ph_by, ph_bz, npx, npy, npz, gn=2): | |
| nbrPoints = npx * npy * npz | |
| b = np.zeros((nbrPoints, 3), dtype="f") | |
| # pure primal arrays | |
| bx = np.zeros((npx, npy, npz), dtype=np.float32) | |
| by = np.zeros((npx, npy, npz), dtype=np.float32) | |
| bz = np.zeros((npx, npy, npz), dtype=np.float32) | |
| domainP1 = slice(gn - 1, -gn + 1) | |
| domain = slice(gn, -gn) | |
| # converts yee to pure primal | |
| # we average in the dual direction so we need one extract ghost data in that direction | |
| if ndim_from(npx, npy, npz) == 2: | |
| bx[:, :, 0] = ( | |
| ph_bx[gn:-gn, gn - 1 : -gn + 1][:, 1:] | |
| + ph_bx[gn:-gn, gn - 1 : -gn + 1][:, :-1] | |
| def ndim_from(npx, npy, npz): | |
| if npx > 2 and npy > 2 and npz > 2: | |
| return 3 | |
| elif npx > 2 and npy > 2 and npz == 2: | |
| return 2 | |
| elif npx > 2 and npy == 2 and npz == 2: | |
| return 1 | |
| else: | |
| raise ValueError( | |
| f" cannot infer dimension from (npx, npy, npz) = {npx} {npy} {npz}" | |
| ) | |
| def BtoFlatPrimal(ph_bx, ph_by, ph_bz, npx, npy, npz, gn=2): | |
| ndim = 2 if npz <= 2 else 3 | |
| nbrPoints = npx * npy * npz | |
| b = np.zeros((nbrPoints, 3), dtype="f") | |
| # pure primal arrays | |
| bx = np.zeros((npx, npy, npz), dtype=np.float32) | |
| by = np.zeros((npx, npy, npz), dtype=np.float32) | |
| bz = np.zeros((npx, npy, npz), dtype=np.float32) | |
| domainP1 = slice(gn - 1, -gn + 1) | |
| domain = slice(gn, -gn) | |
| # converts yee to pure primal | |
| # we average in the dual direction so we need one extract ghost data in that direction | |
| if ndim == 2: | |
| bx[:, :, 0] = ( | |
| ph_bx[gn:-gn, gn - 1 : -gn + 1][:, 1:] + ph_bx[gn:-gn, gn - 1 : -gn + 1][:, :-1] | |
| ) * 0.5 | |
| by[:, :, 0] = ( | |
| ph_by[gn - 1 : -gn + 1, gn:-gn][1:, :] + ph_by[gn - 1 : -gn + 1, gn:-gn][:-1, :] | |
| ) * 0.5 | |
| bz[:, :, 0] = ( | |
| ph_bz[gn - 1 : -gn + 1, gn - 1 : -gn + 1][1:, 1:] | |
| + ph_bz[gn - 1 : -gn + 1, gn - 1 : -gn + 1][:-1, :-1] | |
| ) * 0.5 | |
| # Duplicate 2D data for pseudo-3D representation | |
| for i in range(1, npz): | |
| bx[:, :, i] = bx[:, :, 0] | |
| by[:, :, i] = by[:, :, 0] | |
| bz[:, :, i] = bz[:, :, 0] | |
| else: | |
| # Add 3D handling code here when needed | |
| raise NotImplementedError("3D support not yet implemented") |
| def times_in(phare_h5): | ||
| times_str = list(phare_h5["t"].keys()) | ||
| times = np.asarray([float(time) for time in times_str]) | ||
| times.sort() | ||
| return times | ||
|
|
||
|
|
||
| def is_vector_data(patch): | ||
| return len(patch.keys()) == 3 | ||
|
|
||
|
|
||
| def level_spacing_from(root_spacing, ilvl): | ||
| # hard-coded 2D adds 0 for last dim spacing | ||
| return [dl / 2**ilvl for dl in root_spacing] | ||
|
|
||
|
|
||
| def make3d(root_spacing): | ||
| if len(root_spacing) == 1: | ||
| return root_spacing + [0, 0] | ||
| elif len(root_spacing) == 2: | ||
| return root_spacing + [0] | ||
| return root_spacing | ||
|
|
||
|
|
||
| def main(): | ||
|
|
||
| if len(sys.argv) != 2 or sys.argv[1] in ["-h", "--help"]: | ||
| print(f"Usage: {os.path.basename(sys.argv[0])} <path_to_phare_h5>") | ||
| print("Works for EM fields, bulk velocity and density") | ||
| sys.exit(1) | ||
|
|
||
| path = sys.argv[1] |
There was a problem hiding this comment.
Variable toFlatPrimal may be used uninitialized
If primalFlattener returns a function that doesn't match the expected signature, this can lead to runtime errors.
def primalFlattener(diag_filename):
+ """
+ Returns the appropriate function to convert data to flat primal format.
+
+ The returned function will have one of these signatures:
+ - For vector data: f(ph_x, ph_y, ph_z, npx, npy, npz, gn=2) -> np.array
+ - For scalar data: f(ph_scalar, npx, npy, npz, gn=2) -> np.array
+ """
if "_B" in diag_filename:
print("Converting B fields")
return BtoFlatPrimalAdd validation to ensure proper function type is used:
if is_vector_data(patch):
x_name, y_name, z_name = list(patch.keys())
ph_x = patch[x_name][:]
ph_y = patch[y_name][:]
ph_z = patch[z_name][:]
box = boxFromPatch(patch)
AMRBox.append(box)
nbr_boxes += 1
npx, npy, npz = nbrNodes(box)
+ # Ensure toFlatPrimal is a vector conversion function
+ if toFlatPrimal.__name__ not in ['BtoFlatPrimal', 'EtoFlatPrimal', 'primalVectorToFlatPrimal']:
+ raise TypeError(f"Expected vector conversion function but got {toFlatPrimal.__name__}")
data = toFlatPrimal(ph_x, ph_y, ph_z, npx, npy, npz)
else:
assert len(patch.keys()) == 1
dataset_name = list(patch.keys())[0]
ph_data = patch[dataset_name][:]
box = boxFromPatch(patch)
AMRBox.append(box)
nbr_boxes += 1
npx, npy, npz = nbrNodes(box)
+ # Ensure toFlatPrimal is a scalar conversion function
+ if toFlatPrimal.__name__ != 'primalScalarToFlatPrimal':
+ raise TypeError(f"Expected scalar conversion function but got {toFlatPrimal.__name__}")
data = toFlatPrimal(ph_data, npx, npy, npz)📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| def times_in(phare_h5): | |
| times_str = list(phare_h5["t"].keys()) | |
| times = np.asarray([float(time) for time in times_str]) | |
| times.sort() | |
| return times | |
| def is_vector_data(patch): | |
| return len(patch.keys()) == 3 | |
| def level_spacing_from(root_spacing, ilvl): | |
| # hard-coded 2D adds 0 for last dim spacing | |
| return [dl / 2**ilvl for dl in root_spacing] | |
| def make3d(root_spacing): | |
| if len(root_spacing) == 1: | |
| return root_spacing + [0, 0] | |
| elif len(root_spacing) == 2: | |
| return root_spacing + [0] | |
| return root_spacing | |
| def main(): | |
| if len(sys.argv) != 2 or sys.argv[1] in ["-h", "--help"]: | |
| print(f"Usage: {os.path.basename(sys.argv[0])} <path_to_phare_h5>") | |
| print("Works for EM fields, bulk velocity and density") | |
| sys.exit(1) | |
| path = sys.argv[1] | |
| def times_in(phare_h5): | |
| times_str = list(phare_h5["t"].keys()) | |
| times = np.asarray([float(time) for time in times_str]) | |
| times.sort() | |
| return times | |
| def is_vector_data(patch): | |
| return len(patch.keys()) == 3 | |
| def level_spacing_from(root_spacing, ilvl): | |
| # hard-coded 2D adds 0 for last dim spacing | |
| return [dl / 2**ilvl for dl in root_spacing] | |
| def make3d(root_spacing): | |
| if len(root_spacing) == 1: | |
| return root_spacing + [0, 0] | |
| elif len(root_spacing) == 2: | |
| return root_spacing + [0] | |
| return root_spacing | |
| def main(): | |
| if len(sys.argv) != 2 or sys.argv[1] in ["-h", "--help"]: | |
| print(f"Usage: {os.path.basename(sys.argv[0])} <path_to_phare_h5>") | |
| print("Works for EM fields, bulk velocity and density") | |
| sys.exit(1) | |
| path = sys.argv[1] | |
| # Assume diag_filename is derived from path or another source | |
| toFlatPrimal = primalFlattener(path) | |
| # ... (other code initializing variables like AMRBox, nbr_boxes, etc.) | |
| # Example loop processing patches (the actual context may vary) | |
| for patch in patches: | |
| if is_vector_data(patch): | |
| x_name, y_name, z_name = list(patch.keys()) | |
| ph_x = patch[x_name][:] | |
| ph_y = patch[y_name][:] | |
| ph_z = patch[z_name][:] | |
| box = boxFromPatch(patch) | |
| AMRBox.append(box) | |
| nbr_boxes += 1 | |
| npx, npy, npz = nbrNodes(box) | |
| # Ensure toFlatPrimal is a vector conversion function | |
| if toFlatPrimal.__name__ not in ['BtoFlatPrimal', 'EtoFlatPrimal', 'primalVectorToFlatPrimal']: | |
| raise TypeError(f"Expected vector conversion function but got {toFlatPrimal.__name__}") | |
| data = toFlatPrimal(ph_x, ph_y, ph_z, npx, npy, npz) | |
| else: | |
| assert len(patch.keys()) == 1 | |
| dataset_name = list(patch.keys())[0] | |
| ph_data = patch[dataset_name][:] | |
| box = boxFromPatch(patch) | |
| AMRBox.append(box) | |
| nbr_boxes += 1 | |
| npx, npy, npz = nbrNodes(box) | |
| # Ensure toFlatPrimal is a scalar conversion function | |
| if toFlatPrimal.__name__ != 'primalScalarToFlatPrimal': | |
| raise TypeError(f"Expected scalar conversion function but got {toFlatPrimal.__name__}") | |
| data = toFlatPrimal(ph_data, npx, npy, npz) | |
| # ... (rest of main function) | |
| def primalFlattener(diag_filename): | |
| """ | |
| Returns the appropriate function to convert data to flat primal format. | |
| The returned function will have one of these signatures: | |
| - For vector data: f(ph_x, ph_y, ph_z, npx, npy, npz, gn=2) -> np.array | |
| - For scalar data: f(ph_scalar, npx, npy, npz, gn=2) -> np.array | |
| """ | |
| if "_B" in diag_filename: | |
| print("Converting B fields") | |
| return BtoFlatPrimal | |
| # Additional logic may follow to select and return the correct conversion function |
Just a quick and incomplete (just handles E and B) script to convert our data to vtkhdf and open it with paraview.
Just here as a way to learn paraview and decide whether it's a viable way for future diag formats.