Merge pull request #265 from apax-hub/spell-check

spell check

apax/bal/api.py

@@ -117,9 +117,9 @@ def kernel_selection(
     selection_method:
         Currently only "max_dist" is supported.
     feature_transforms:
-        Feature tranforms to be applied on top of the
+        Feature transforms to be applied on top of the
         base feature map transform.
-        Examples would include multiplcation with or addition of a constant.
+        Examples would include multiplication with or addition of a constant.
     selection_batch_size:
         Amount of new data points to be selected from `pool_atoms`.
     processing_batch_size:

apax/config/train_config.py

@@ -359,7 +359,7 @@ class CheckpointConfig(BaseModel, extra="forbid"):
 
 class Config(BaseModel, frozen=True, extra="forbid"):
     """
-    Main configuration of a apax training run. Parameter that are cofig classes will
+    Main configuration of a apax training run. Parameter that are config classes will
     be generated by parsing the config.yaml file and are specified
     as shown :ref:`here <train_config>`:
 
@@ -385,7 +385,7 @@ class Config(BaseModel, frozen=True, extra="forbid"):
         | Number of models to be trained at once.
     n_jitted_steps : int, default = 1
         | Number of train batches to be processed in a compiled loop.
-        | Can yield singificant speedups for small structures or small batch sizes.
+        | Can yield significant speedups for small structures or small batch sizes.
     data : :class:`.DataConfig`
         | Data configuration.
     model : :class:`.ModelConfig`

apax/data/input_pipeline.py

@@ -18,7 +18,7 @@
 
 def pad_nl(idx, offsets, max_neighbors):
     """
-    Pad the neighbor list arrays to the maximal number of neighbors occuring.
+    Pad the neighbor list arrays to the maximal number of neighbors occurring.
 
     Parameters
     ----------

apax/md/ase_calc.py

@@ -305,7 +305,7 @@ def batch_eval(
             unpadded_results = unpack_results(results, inputs)
 
             # for the last batch, the number of structures may be less
-            # than the batch_size,  which is why we check this explicitely
+            # than the batch_size,  which is why we check this explicitly
             num_strucutres_in_batch = results["energy"].shape[0]
             for j in range(num_strucutres_in_batch):
                 atoms = atoms_list[i].copy()

apax/train/run.py

@@ -139,7 +139,7 @@ def run(user_config: Union[str, os.PathLike, dict], log_level="error"):
     Parameters
     ----------
     user_config : str | os.PathLike | dict
-        training config full exmaple can be finde :ref:`here <train_config>`:
+        training config full example can be find :ref:`here <train_config>`:
 
     """
     config = parse_config(user_config)

apax/utils/jax_md_reduced/partition.py

@@ -161,7 +161,7 @@ def update(self, position: Array, **kwargs) -> "CellList":
     def kwarg_buffers(self):
         logging.warning(
             "kwarg_buffers renamed to named_buffer. The name "
-            "kwarg_buffers will be depricated."
+            "kwarg_buffers will be deprecated."
         )
         return self.named_buffer
 
@@ -179,7 +179,7 @@ class PartitionErrorCode(IntEnum):
         to allocate a new cell list.
       CELL_SIZE_TOO_SMALL: Indicates that the size of cells in a cell list was
         not large enough to properly capture particle interactions. This
-        indicates that it is necessary to allcoate a new cell list with larger
+        indicates that it is necessary to allocate a new cell list with larger
         cells.
       MALFORMED_BOX: Indicates that a box matrix was not properly upper
         triangular.
@@ -242,7 +242,7 @@ class NeighborList:
       reference_position: The positions of particles when the neighbor list was
         constructed. This is used to decide whether the neighbor list ought to be
         updated.
-      error: An error code that is used to identify errors that occured during
+      error: An error code that is used to identify errors that occurred during
         neighbor list construction. See `PartitionError` and `PartitionErrorCode`
         for details.
       cell_list_capacity: An optional integer specifying the capacity of the cell
@@ -317,7 +317,7 @@ class NeighborList:
       reference_position: The positions of particles when the neighbor list was
         constructed. This is used to decide whether the neighbor list ought to be
         updated.
-      error: An error code that is used to identify errors that occured during
+      error: An error code that is used to identify errors that occurred during
         neighbor list construction. See `PartitionError` and `PartitionErrorCode`
         for details.
       cell_list_capacity: An optional integer specifying the capacity of the cell

apax/utils/jax_md_reduced/simulate.py

@@ -1590,8 +1590,8 @@ def temp_csvr(
 
     Samples from the canonical ensemble in which the number of particles (N),
     the system volume (V), and the temperature (T) are held constant. CSVR
-    algorithmn samples the canonical distribution by rescaling the velocities
-    by a appropritely chosen random factor. At each timestep (dt) the rescaling
+    algorithm samples the canonical distribution by rescaling the velocities
+    by a appropriately chosen random factor. At each timestep (dt) the rescaling
     takes place and the rescaling factor is calculated using
     A7 Bussi et al. [#bussi2007]_. CSVR updates to the velocity are stochastic in
     nature and unlike the Berendsen thermostat it samples the true canonical

apax/utils/jax_md_reduced/smap.py

@@ -160,7 +160,7 @@ def bond(
         an ndarray of distances or displacements of shape `[]` or `[d_in]`
         respectively. The metric can optionally take a floating point time as a
         third argument.
-      static_bonds: An ndarray of integer pairs wth shape `[b, 2]` where each
+      static_bonds: An ndarray of integer pairs with shape `[b, 2]` where each
         pair specifies a bond. `static_bonds` are baked into the returned compute
         function statically and cannot be changed after the fact.
       static_bond_types: An ndarray of integers of shape `[b]` specifying the

examples/01_Model_Training.ipynb

@@ -21,7 +21,7 @@
     "\n",
     "## Acquiring a dataset\n",
     "\n",
-    "You can obtain the benzene dataset with DFT labels either by running the following command or manually from this [link](http://www.quantum-machine.org/gdml/data/xyz/ethanol_ccsd_t.zip). Apax uses ASE to read in datasets, so make sure to convert your own data into an ASE readable format (extxyz, traj etc). Be carefull the downloaded dataset has to be modified like in the `apax.untils.dataset.mod_md_datasets` function in order to be readable."
+    "You can obtain the benzene dataset with DFT labels either by running the following command or manually from this [link](http://www.quantum-machine.org/gdml/data/xyz/ethanol_ccsd_t.zip). Apax uses ASE to read in datasets, so make sure to convert your own data into an ASE readable format (extxyz, traj etc). Be careful the downloaded dataset has to be modified like in the `apax.utils.dataset.mod_md_datasets` function in order to be readable."
    ]
   },
   {
@@ -100,7 +100,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "The following command create a minimal configuration file in the working directory. Full configuration file with descriptiond of the prameter can be found [here](https://github.com/apax-hub/apax/blob/main/apax/cli/templates/train_config_full.yaml)."
+    "The following command create a minimal configuration file in the working directory. Full configuration file with descriptiond of the parameter can be found [here](https://github.com/apax-hub/apax/blob/main/apax/cli/templates/train_config_full.yaml)."
    ]
   },
   {
@@ -291,7 +291,7 @@
     "\n",
     "If training is interrupted for any reason, re-running the above `train` command will resume training from the latest checkpoint.\n",
     "\n",
-    "Furthermore, an Apax trianing can easily be started within a script."
+    "Furthermore, an Apax training can easily be started within a script."
    ]
   },
   {

examples/02_Molecular_Dynamics.ipynb

@@ -244,7 +244,7 @@
     "duration: 20_000 # fs\n",
     "initial_structure: project/benzene_mod.xyz\n",
     "```\n",
-    "Full configuration file with descriptiond of the prameter can be found [here](https://github.com/apax-hub/apax/blob/main/apax/cli/templates/md_config_minimal.yaml)."
+    "Full configuration file with descriptiond of the parameter can be found [here](https://github.com/apax-hub/apax/blob/main/apax/cli/templates/md_config_minimal.yaml)."
    ]
   },
   {
@@ -339,7 +339,7 @@
    "metadata": {},
    "source": [
     "During the simulation, a progress bar tracks the instantaneous temperature at each outer step.\n",
-    "The simulation is followd by a small oh bondlength distribution analyses of the trajectory defined [here](#bondlength)."
+    "The simulation is followed by a small oh bondlength distribution analyses of the trajectory defined [here](#bondlength)."
    ]
   },
   {

examples/04_Batch_Data_Selection.ipynb

@@ -7,7 +7,7 @@
     "# Batch Active Learning\n",
     "\n",
     "While it is possible to perform rudimentary data selection simply by randomly choosing samples, the batch of data thus drawn might not be the most informative one.\n",
-    "Choosing those samples whith the largest prediction uncertainties from trajectories often results in the selection of configurations from subsequent time steps.\n",
+    "Choosing those samples with the largest prediction uncertainties from trajectories often results in the selection of configurations from subsequent time steps.\n",
     "\n",
     "Batch selection methods can be constructed to select informative and diverse data, with or without following the underlying distribution.\n",
     "\n",
@@ -392,7 +392,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "As we can see below, the batch selection method only picks a few data points from the optimization part of the pool, indicating that during an optmization the structure of the molecule does not change very much.\n",
+    "As we can see below, the batch selection method only picks a few data points from the optimization part of the pool, indicating that during an optimization the structure of the molecule does not change very much.\n",
     "Hence, there are not many informative samples to be found in it."
    ]
   },

tests/unit_tests/data/test_input_pipeline.py

@@ -10,7 +10,7 @@
 from apax.utils.data import split_atoms, split_idxs
 from apax.utils.random import seed_py_np_tf
 
-# TODO REENABLE LATER
+# TODO RE-ENABLE LATER
 # @pytest.mark.parametrize(
 #     "num_data, pbc, calc_results, external_labels",
 #     (