Change polarizations and q-vectors convention

Cartesian coordinates are used to indicate polarization of
incoming and outgoing photons, as well as for q-vectors.

docs/source/howto/overrides.md

@@ -1,6 +1,6 @@
 (howto-overrides)=
 
-# How-to use protocols and overrides
+# Protocols and overrides
 
 :::{important}
 The following how-to assume that you are familiar with submitting the workflows of the package.
@@ -17,7 +17,7 @@ as you probably noticed in the [tutorials](tutorials).
 Not necesseraly all the WorkChains have the `get_builder_from_protocol`. E.g. The {class}`~aiida_vibroscopy.workflows.dielectric.numerical_derivatives`.
 :::
 
-## How to use in general
+## General usage
 
 Depending on the WorkChain, you will need to specify some minimal, _but required_, inputs. Here is an example:
 
@@ -48,7 +48,7 @@ Out[1]:
 
 ```
 
-## How to use overrides (beginner)
+## Overrides (beginner)
 
 As stated at the beginning, you might need to tweak your builder inputs before submitting. Instead of modifying the inputs via the builder, you can do it via `overrides`. The way the overrides must be secified __is WorkChain dependent__. The structure should be the same as the input specification of the specific WorkChain, so always refer to the inputs of the particular workflow, which you can find [here](topics-workflows).
 
@@ -169,7 +169,7 @@ builder = DielectricWorkChain.get_builder_from_protocol(code=code, structure=str
 
 You got the gist, so it will be the same for the `phonon` inputs, which corresponds to the `PhononWorkChain`.
 
-## How to overrides (advanced)
+## Overrides (advanced)
 
 Once you got used to using the overrides, you will notice your python script will become quite a mess.
 The idea is to write the overrides in a separate file, using for example the convenient `YAML` format,
@@ -199,7 +199,7 @@ phonon:
     scf:
         pw:
             kpoints_distance: 0.2
-            pw:
+            parameters:
                 ...
 dielectric:
     central_difference:
@@ -209,7 +209,7 @@ dielectric:
             ...
 ```
 
-## Get automated inputs for insulators, metals, and magnetism
+## Automated inputs for insulators, metals, and magnetism
 
 Inputs might change depending on the nature of the material: insulating, metallic, a form of magnetism.
 This can be the need of specifying a smearing or a magnetic configuration.

docs/source/howto/postprocess.md

@@ -1,6 +1,6 @@
 (howto-postprocess)=
 
-# How-to post-process data
+# Post-process data
 
 Here you will find some _how-tos_ on how to post-process the `VibrationalData` and `PhonopyData`.
 These can be generated via the `HarmonicWorkChain`, the `IRamanSpectraWorkChain` or the `PhononWorkChain`.
@@ -10,13 +10,12 @@ Please, have a look at the [tutorials](tutorials) for an overview, and at the sp
 You can always access to the detailed description of a function/method putting a `?` question mark in front
 of the function and press enter. This will print the description of the function, with detailed information
 on inputs, outputs and purpose.
-``` {note}
+
 A rendered version can also be found in the documentation referring to
 the dedicated [API section](reference).
-```
 :::
 
-## How to get powder spectra
+## Powder spectra
 
 You can get the powder infrared and/or Raman spectra from a computed {{ vibrational_data }}. For Raman:
 
@@ -28,8 +27,8 @@ vibro = load_node(IDENTIFIER) # a VibrationalData node
 polarized_intensities, depolarized_intensities, frequencies, labels = vibro.run_powder_raman_intensities(frequency_laser=532, temperature=300)
 ```
 
-The total powder intensity will be the some of the polarized (backscattering geometry)
-and depolarized (90º geometry) intensities:
+The total powder intensity will be the some of the polarized (backscattering geometry, HH/VV setup)
+and depolarized (90º geometry, HV setup) intensities:
 
 ```python
 total = polarized_intensities + depolarized_intensities
@@ -44,7 +43,7 @@ intensities, frequencies, labels = vibro.run_powder_ir_intensities()
 The `labels` output is referring to the irreducible representation labels of the modes.
 
 
-## How to get single crystal spectra
+## Single crystal polarized spectra
 
 ::: {important}
 It is extremely important that you match the experimental conditions, meaning the direction of
@@ -53,7 +52,7 @@ on the convention used, both in the code and in the experiments.
 :::
 
 ::: {note} Convention
-In the following methods, the direction of the light must be given in crystal/fractional coordinates.
+In the following methods, the direction of the light must be given in **Cartesian coordinates**.
 :::
 
 You can get the single crystal infrared and/or Raman spectra from a computed {{ vibrational_data }}. For Raman:
@@ -86,7 +85,7 @@ incoming = [0,0,1] # light polatization of the incident laser beam
 intensities, frequencies, labels = vibro.run_single_crystal_ir_intensities(pol_incoming=incoming)
 ```
 
-::: {admonition} Cartesian coordinates
+<!-- ::: {admonition} Cartesian coordinates
 :class: hint
 To get the direction in Cartesian coordinates expressed in crystal coordinates of the primitive cell, you can use the following
 snippet
@@ -99,14 +98,9 @@ incoming_cartesian = [0,0,1]
 inv_cell = np.linalg.inv(cell)
 incoming = np.dot(invcell.T, incoming_cartesian)
 ```
-For the q-direction instead you need to transform them into reciprocal space crystal coordinates, as follows
-```python
-q_nac_cartesian = [0,0,1]
-q_nac_crystal = np.dot(cell, q_nac_cartesian)
-```
-:::
+::: -->
 
-## How to get the IR/Raman active modes
+## IR/Raman active modes
 
 To get the active modes from a computed {{ vibrational_data }}:
 

docs/source/howto/understand.md

@@ -1,6 +1,6 @@
 (howto-understand)=
 
-# How-to understand the input/builder structure
+# Understand the input/builder structure
 
 In AiiDA the CalcJobs and WorkChains have usually nested inputs and different options on how to run the calculation
 and/or workflows. To understand the nested input structure of CalcJobs/Workflows, we made them available in a clickable

docs/source/topics/conventions.md

@@ -15,7 +15,7 @@ parenthesis we give the corresponding attributes to access quantity via the {{ v
     3. Atomic displacements index.
 
 - Non-linear optical susceptibility (`nlo_susceptibility`): in units of pm/V; 3rd rank tensors, with equivalent indices.
-- Raman tensors (`raman_tensors`): in units of 4pi/Angstrom; array with 4 indices, respectively:
+- Raman tensors (`raman_tensors`): in units of 1/Angstrom; array with 4 indices, respectively:
 
     1. Atomic index.
     2. Atomic displacements index.

src/aiida_vibroscopy/calculations/numerical_derivatives_utils.py

@@ -194,11 +194,11 @@ def compute_susceptibility_derivatives(
     preprocess_data: PreProcessData, electric_field: orm.Float, diagonal_scale: orm.Float, accuracy_order: orm.Int,
     **kwargs
 ) -> dict:
-    """Return the Raman (4pi/Angstrom) and the non-linear optical susceptibility (pm/V) tensors.
+    """Return the Raman (1/Angstrom) and the non-linear optical susceptibility (pm/V) tensors.
 
     ..note:
         * If the numerical accuracy order is greater than 2, arrays at lower orders are given as well.
-        * Units are 4pi/Angstrom for Raman tensors, normalized using the UNITCELL volume.
+        * Units are 1/Angstrom for Raman tensors, normalized using the UNITCELL volume.
         * Units are pm/V for non-linear optical susceptibility
         * Raman tensors indecis: (atomic,  atomic displacement, electric field, electric field)
 
@@ -378,7 +378,7 @@ def compute_susceptibility_derivatives(
         chis_data.update({key_order: deepcopy(chis_array_data)})
 
     units_data = orm.Dict({
-        'raman_tensors': r'$4\pi/\AA$',
+        'raman_tensors': r'$1/\AA$',
         'nlo_susceptibility': 'pm/V',
     })
 

src/aiida_vibroscopy/calculations/spectra_utils.py

@@ -54,7 +54,7 @@ def compute_active_modes(
     Raman and infrared active modes can be extracted using `selection_rule`.
 
     :param nac_direction: (3,) shape list, indicating non analytical
-        direction in fractional unit cell space coordinates
+        direction in Cartesian coordinates
     :param selection_rule: str, can be `raman` or `ir`, it uses symmetry in
         the selection of the modes for a specific type of process
     :param sr_thr: float, threshold for selection rule (the analytical value is 0)
@@ -69,9 +69,11 @@ def compute_active_modes(
     nac_dir = [0, 0, 0] if nac_direction is None else np.array(nac_direction)
 
     # Notation: columns convention; k = reciprocal; x = direct space; (p) = primitive; (u) = unitcell
-    # C(p) = C(u) * M --> M = C(u)^-1 * C(p)
+    # C(p) = C(u) * M  ==> M = C(u)^-1 * C(p)
     # x(p) = M^-1 x(u) ==> k(p) = M^T k(u);
-    q_reduced = np.dot(phonopy_instance.primitive_matrix.T, nac_dir)  # in reduced (PRIMITIVE) coordinates
+    # q_reduced = np.dot(phonopy_instance.primitive_matrix.T, nac_dir)  # in reduced/crystal (PRIMITIVE) coordinates
+    q_reduced = np.dot(phonopy_instance.primitive.cell,
+                       nac_dir) / (2. * np.pi)  # in reduced/crystal (PRIMITIVE) coordinates
 
     # Step 1 - set the irreducible representations and the phonons
     phonopy_instance.set_irreps(q=[0, 0, 0], nac_q_direction=q_reduced, degeneracy_tolerance=degeneracy_tolerance)
@@ -191,8 +193,8 @@ def compute_raman_susceptibility_tensors(
         * Unitcell volume for Raman tensor as normalization (in case non-primitive cell was used).
 
     :param phonopy_instance: Phonopy instance with non-analytical constants included
-    :param nac_direction: non-analytical direction in unit cell space coordinates
-    :param raman_tensors: dChi/du in Cartesian coordinates (in 4pi/Angstrom)
+    :param nac_direction: non-analytical direction in Cartesian coordinates
+    :param raman_tensors: dChi/du in Cartesian coordinates (in 1/Angstrom)
     :param nlo_susceptibility: non linear optical susceptibility
         in Cartesian coordinates (in pm/V)
     :param use_irreps: whether to use irreducible representations
@@ -213,8 +215,9 @@ def compute_raman_susceptibility_tensors(
     sqrt_volume = np.sqrt(volume)
     raman_tensors *= volume
 
-    ucell = phonopy_instance.unitcell.cell.T  # as rows
-    q_cartesian = np.dot(ucell, nac_direction)  # in Cartesian coordinates
+    # rcell = np.linalg.inv(phonopy_instance.unitcell.cell) # as columns
+    # q_cartesian = np.dot(rcell, nac_direction)  # in Cartesian coordinates
+    q_cartesian = nac_direction
 
     selection_rule = 'raman' if use_irreps else None
 
@@ -303,7 +306,7 @@ def compute_polarization_vectors(
     .. note:: the unite for polarization vectors are in (debey/angstrom)/sqrt(AMU)
 
     :param phonopy_instance: Phonopy instance with non-analytical constants included
-    :param nac_direction: non-analytical direction in fractional coordinates (unit cell)
+    :param nac_direction: non-analytical direction in Cartesian coordinates
     :param use_irreps: whether to use irreducible representations
         in the selection of modes, defaults to True
     :param degeneracy_tolerance: degeneracy tolerance
@@ -355,7 +358,7 @@ def compute_complex_dielectric(
     :param freq_range: frequency range in cm^-1; set to `auto` for automatic choice
     :param gammas: list or single value of broadenings, i.e. full width at half maximum (FWHM)
     :param nac_direction: (3,) shape list, indicating non analytical
-        direction in fractional unit cell space coordinates
+        direction in Cartesian coordinates
     :param use_irreps: whether to use irreducible representations
         in the selection of modes, defaults to True
     :param degeneracy_tolerance: degeneracy tolerance
@@ -387,7 +390,7 @@ def compute_complex_dielectric(
     else:
         sigmas = [float(gammas) for _ in frequencies]
 
-    if freq_range == 'auto':
+    if isinstance(freq_range, str):
         xi = max(0, frequencies.min() - 200)
         xf = frequencies.max() + 200
         freq_range = np.arange(xi, xf, 1.)