chore: image filters for better dark mode contrast

.github/workflows/deploy-gh-pages.yml

@@ -26,6 +26,9 @@ jobs:
       - name: Install NPM packages
         run: npm install
 
+      - name: Apply customizations
+        run: sed -i 's/rgba(239,\s*83,\s*80,\s*0.56)/rgba(239, 83, 80, 0.9)/gI' node_modules/prism-react-renderer/dist/index.js
+
       - name: Build website
         run: npm run build
 

docs/hands-on/Bi2Se3.md

@@ -44,7 +44,7 @@ spin-orbit calculation without spin magnetization. It assumes that time reversal
 symmetry holds and it does not calculate the magnetization. The states are
 still two-component spinors but the total magnetization is zero.
 
-![Bi2Se3-bands](../../static/img/Bi2Se3-bands.webp)
+<img src={require("../../static/img/Bi2Se3-bands.webp").default} class="inv-hue-rot-180" alt="Bi2Se3-bands"/>
 
 Notice that for the Dirac surface states the gap did not completely close at the
 Fermi energy. This is possibly due to finite size effect. We could repeat the
@@ -55,7 +55,7 @@ In order to sample the $\Gamma$ point for our DOS calculation, an odd k-grid
 mesh (25✕25✕5) was used. The signature of Dirac cone is evident from the DOS
 figure.
 
-![Bi2Se3-dos](../../static/img/Bi2Se3-dos.webp)
+<img src={require("../../static/img/Bi2Se3-dos.webp").default} class="inv-hue-rot-180" alt="Bi2Se3-dos"/>
 
 ## Resources
 - [https://docs.quantumatk.com/tutorials/topological_insulator_bi2se3/](

docs/hands-on/GaAs.md

@@ -38,7 +38,7 @@ bands.x < pp.bands.GaAs.in > pp.bands.GaAs.out
 ```
 If everything goes well, you will get the bandstructure as below:
 
-![GaAs-bands](../../static/img/GaAs-bands.webp)
+<img src={require("../../static/img/GaAs-bands.webp").default} class="inv-hue-rot-180" alt="GaAs-bands"/>
 
 :::caution Warning
 

docs/hands-on/aluminum.mdx

@@ -66,7 +66,7 @@ dos.x < al_dos.in > al_dos.out
 Note from our `al_nscf.out` that our Fermi energy is at 7.9421 eV. We plot our
 density of states:
 
-![al-dos](../../static/img/al-dos.webp)
+<img src={require("../../static/img/al-dos.webp").default} class="inv-hue-rot-180" alt="al-dos"/>
 
 ## Bandstructure calculation
 
@@ -93,7 +93,7 @@ bands.x < al_bands_pp.in > al_bands_pp.out
 ```
 We obtain the following bandstructure:
 
-![al-bands](../../static/img/al-bands.webp)
+<img src={require("../../static/img/al-bands.webp").default} class="inv-hue-rot-180" alt="al-bands"/>
 
 ## Importance of smearing in convergence
 
@@ -115,7 +115,7 @@ Marzari-Vanderbilt), and for various smearing values.
 pwtk al.degauss.pwtk
 ```
 
-![al-smearing](../../static/img/al-smearing.webp)
+<img src={require("../../static/img/al-smearing.webp").default} class="inv-hue-rot-180" alt="al-smearing"/>
 
 We see that the `m-v` and `m-p` broadening allow for faster and smother
 convergence while depending less on `degauss` value than Gaussian broadening.

docs/hands-on/bands.mdx

@@ -118,7 +118,7 @@ plt.text(2.3, 5.6, 'Fermi energy', fontsize= small)
 plt.show()
 ```
 
-![silicon-bands](../../static/img/silicon-bands.webp)
+<img src={require("../../static/img/silicon-bands.webp").default} class="inv-hue-rot-180" alt="silicon-bands"/>
 
 :::info
 

docs/hands-on/convergence.mdx

@@ -41,7 +41,7 @@ plt.legend(frameon=False)
 plt.show()
 ```
 
-![etot-vs-ecutwfc](../../static/img/etot-vs-ecutwfc.webp)
+<img src={require("../../static/img/etot-vs-ecutwfc.webp").default} class="inv-hue-rot-180" alt="etot-vs-ecutwfc"/>
 
 ## Convergence test using UNIX shell script
 
@@ -106,7 +106,7 @@ plt.legend(frameon=False)
 plt.show()
 ```
 
-![etot-vs-kpoints](../../static/img/etot-vs-kpoints.webp)
+<img src={require("../../static/img/etot-vs-kpoints.webp").default} class="inv-hue-rot-180" alt="etot-vs-kpoints"/>
 
 ## Convergence against lattice constant
 
@@ -131,7 +131,7 @@ plt.legend(frameon=False)
 plt.show()
 ```
 
-![etot-vs-alat](../../static/img/etot-vs-alat.webp)
+<img src={require("../../static/img/etot-vs-alat.webp").default} class="inv-hue-rot-180" alt="etot-vs-alat"/>
 
 
 ## Note on CPU time

docs/hands-on/dft-u.mdx

@@ -86,7 +86,7 @@ pw.x -in feo_nscf.in > feo_nscf.out
 projwfc.x -in feo_projwfc.in > feo_projwfc.out
 ```
 
-![feo-pdos-dft](../../static/img/feo-pdos-dft.webp)
+<img src={require("../../static/img/feo-pdos-dft.webp").default} class="inv-hue-rot-180" alt="feo-pdos-dft"/>
 
 This gives us metallic density of states. In practice we get insulating FeO.
 
@@ -144,7 +144,7 @@ Hubbard_U(2) = 4.6
 We repeat the above calculation and plot the results. Now we find insulating
 ground state.
 
-![feo-pdos-dft-u](../../static/img/feo-pdos-dft-u.webp)
+<img src={require("../../static/img/feo-pdos-dft-u.webp").default} class="inv-hue-rot-180" alt="feo-pdos-dft-u"/>
 
 :::info
 

docs/hands-on/dos.mdx

@@ -89,7 +89,7 @@ plt.text(6, 1.7, 'Fermi energy', fontsize= med, rotation=90)
 plt.show()
 ```
 
-![silicon-dos](../../static/img/silicon-dos.webp)
+<img src={require("../../static/img/silicon-dos.webp").default} class="inv-hue-rot-180" alt="silicon-dos"/>
 
 :::info Important
 

docs/hands-on/epsilon.mdx

@@ -73,8 +73,7 @@ plt.ylabel("$\\epsilon_1~/~\\epsilon_2$")
 plt.legend(frameon=False)
 plt.show()
 ```
-
-![silicon-epsilon](../../static/img/silicon-epsilon.webp)
+<img src={require("../../static/img/silicon-epsilon.webp").default} class="inv-hue-rot-180" alt="silicon-epsilon"/>
 
 :::warning
 

docs/hands-on/fe.mdx

@@ -68,7 +68,7 @@ Run the script:
 pwtk fe_ecut.pwtk
 ```
 
-![fe-convergence](../../static/img/fe-convergence.webp)
+<img src={require("../../static/img/fe-convergence.webp").default} class="inv-hue-rot-180" alt="fe-convergence"/>
 
 ## Density of states calculation
 
@@ -80,9 +80,9 @@ import fe_dos_pwtk from '!!raw-loader!/src/fe/fe_dos.pwtk';
 
 Below is the plots of total and projected density of states.
 
-![fe-dos](../../static/img/fe-dos.webp)
+<img src={require("../../static/img/fe-dos.webp").default} class="inv-hue-rot-180" alt="fe-dos"/>
 
-![fe-pdos](../../static/img/fe-pdos.webp)
+<img src={require("../../static/img/fe-pdos.webp").default} class="inv-hue-rot-180" alt="fe-pdos"/>
 
 Also see bandstructure of Fe with and without [SOC](
 soc#bandstructure-of-fe-with-soc).

docs/hands-on/graphene.md

@@ -26,7 +26,7 @@ pw.x -i graphene_nscf.in > graphene_nscf.out
 dos.x -i graphene_dos.in > graphene_dos.out
 ```
 
-![graphene-dos](../../static/img/graphene-dos.webp)
+<img src={require("../../static/img/graphene-dos.webp").default} class="inv-hue-rot-180" alt="graphene-dos"/>
 
 ## Bandstructure calculation
 
@@ -65,4 +65,4 @@ plt.ylabel("Energy (eV)")
 plt.show()
 ```
 
-![graphene-bands](../../static/img/graphene-bands.webp)
+<img src={require("../../static/img/graphene-bands.webp").default} class="inv-hue-rot-180" alt="graphene-bands"/>

docs/hands-on/ni.mdx

@@ -32,4 +32,4 @@ mpirun -np 8 bands.x -i bands_ni_up.in > bands_ni_up.out
 
 Similarly, we process the spin down bands `spin_component=2` and plot them.
 
-![ni-spin-bands](../../static/img/ni-spin-bands.webp)
+<img src={require("../../static/img/ni-spin-bands.webp").default} class="inv-hue-rot-180" alt="ni-spin-bands"/>

docs/hands-on/pdos.mdx

@@ -100,7 +100,7 @@ plt.show()
 
 Here is how our projected density of states plot looks like:
 
-![al-pdos](../../static/img/al-pdos.webp)
+<img src={require("../../static/img/al-pdos.webp").default} class="inv-hue-rot-180" alt="al-pdos"/>
 
 We can perform sums of specific atom or orbital contributions using
 **sumpdos.x** code if there are multiple $s$ or $p$ orbitals:

docs/hands-on/phonon.mdx

@@ -104,7 +104,7 @@ plt.ylim(0, )
 plt.show()
 ```
 
-![GaAs-phonon](../../static/img/GaAs-phonon.webp)
+<img src={require("../../static/img/GaAs-phonon.webp").default} class="inv-hue-rot-180" alt="GaAs-phonon"/>
 
 :::tip
 
@@ -136,7 +136,7 @@ plt.xlim(freq[0], freq[-1])
 plt.show()
 ```
 
-![GaAs-phonon-dos](../../static/img/GaAs-phonon-dos.webp)
+<img src={require("../../static/img/GaAs-phonon-dos.webp").default} class="inv-hue-rot-180" alt="GaAs-phonon-dos"/>
 
 ## Resources
 

docs/hands-on/soc.mdx

@@ -139,7 +139,7 @@ the file `fe.noncolin.data.3`. All values in this case are either +1/2 or -1/2.
 mpirun -np 8 bands.x -i pp.bands.fe_soc.in > pp.bands.fe_soc.out
 ```
 
-![fe-soc-bands](../../static/img/fe-soc-bands.webp)
+<img src={require("../../static/img/fe-soc-bands.webp").default} class="inv-hue-rot-180" alt="fe-soc-bands"/>
 
 ## SOC calculation for GaAs
 
@@ -152,4 +152,4 @@ mpirun -np 8 pw.x -i pw.bands.GaAs_soc.in > pw.bands.GaAs_soc.out
 mpirun -np 8 bands.x -i pp.bands.GaAs_soc.in > pp.bands.GaAs_soc.out
 ```
 
-![GaAs-soc-bands](../../static/img/GaAs-soc-bands.webp)
+<img src={require("../../static/img/GaAs-soc-bands.webp").default} class="inv-hue-rot-180" alt="GaAs-soc-bands"/>

docs/hands-on/wannier.mdx

@@ -1,60 +1,60 @@
----
-title: Wannier method
----
-
-## Obtain bandstructure of Silicon
-
-1. Perform `scf` calculation using Quantum Espresso `pw.x`
-
-```bash
-QE_PATH="/workspaces/q-e-qe-7.2/bin"
-mpirun -np 4 ${QE_PATH}/pw.x -i pw.scf.silicon.in > pw.scf.silicon.out
-```
-
-2. Perform `nscf` calculation using `pw.x`. Instead of automatic k-grid, we need
-to provide explicit list of k-points. Such explicit list of k-points can be
-generated using perl script included in the Wannier package under utility.
-
-```bash
-WANNIER_PATH="/workspaces/wannier90-3.1.0"
-# directly append the k-points to the input file
-${WANNIER_PATH}/utility/kmesh.pl 4 4 4 >> pw.nscf.silicon.in
-```
-
-Run `nscf` calculation:
-
-```bash
-mpirun -np 4 ${QE_PATH}/pw.x -i pw.nscf.silicon.in > pw.nscf.silicon.out
-```
-
-3. Prepare input file for wannier90 (`silicon.win`). Here we need the k-points
-list without the weights:
-
-```bash
-${WANNIER_PATH}/utility/kmesh.pl 4 4 4 wan
-```
-
-4. Generate nnkp input:
-
-```bash
-# we can just provide the seedname or seedname.win
-${WANNIER_PATH}/wannier90.x -pp silicon
-```
-
-5. Create input file for `pw2wan`, and generate initial projections:
-
-```bash
-mpirun -np 4 ${WANNIER_PATH}/pw2wannier90.x -i pw2wan.silicon.in > pw2wan.silicon.out
-```
-
-6. Run wannier calculation:
-
-```bash
-mpirun -np 4 ${WANNIER_PATH}/wannier90.x silicon
-```
-
-![silicon-wannier-bands](../../static/img/silicon-wannier-bands.webp)
-
-## Resources
-
-- https://sites.google.com/view/hubbard-koopmans/program
+---
+title: Wannier method
+---
+
+## Obtain bandstructure of Silicon
+
+1. Perform `scf` calculation using Quantum Espresso `pw.x`
+
+```bash
+QE_PATH="/workspaces/q-e-qe-7.2/bin"
+mpirun -np 4 ${QE_PATH}/pw.x -i pw.scf.silicon.in > pw.scf.silicon.out
+```
+
+2. Perform `nscf` calculation using `pw.x`. Instead of automatic k-grid, we need
+to provide explicit list of k-points. Such explicit list of k-points can be
+generated using perl script included in the Wannier package under utility.
+
+```bash
+WANNIER_PATH="/workspaces/wannier90-3.1.0"
+# directly append the k-points to the input file
+${WANNIER_PATH}/utility/kmesh.pl 4 4 4 >> pw.nscf.silicon.in
+```
+
+Run `nscf` calculation:
+
+```bash
+mpirun -np 4 ${QE_PATH}/pw.x -i pw.nscf.silicon.in > pw.nscf.silicon.out
+```
+
+3. Prepare input file for wannier90 (`silicon.win`). Here we need the k-points
+list without the weights:
+
+```bash
+${WANNIER_PATH}/utility/kmesh.pl 4 4 4 wan
+```
+
+4. Generate nnkp input:
+
+```bash
+# we can just provide the seedname or seedname.win
+${WANNIER_PATH}/wannier90.x -pp silicon
+```
+
+5. Create input file for `pw2wan`, and generate initial projections:
+
+```bash
+mpirun -np 4 ${WANNIER_PATH}/pw2wannier90.x -i pw2wan.silicon.in > pw2wan.silicon.out
+```
+
+6. Run wannier calculation:
+
+```bash
+mpirun -np 4 ${WANNIER_PATH}/wannier90.x silicon
+```
+
+<img src={require("../../static/img/silicon-wannier-bands.webp").default} class="inv-hue-rot-180" alt="silicon-wannier-bands"/>
+
+## Resources
+
+- https://sites.google.com/view/hubbard-koopmans/program

docs/setup/hpc.mdx

@@ -189,7 +189,7 @@ For my case, the link line was:
 -L${MKLROOT}/lib/intel64 -lmkl_scalapack_lp64 -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -lmkl_blacs_intelmpi_lp64 -lpthread -lm -ldl
 ```
 
-<img src={require("../../static/img/intel-link-line-adviser.webp").default} alt="intel-link-line-adviser" width="600px" />
+<img src={require("../../static/img/intel-link-line-adviser.webp").default} class="inv-hue-rot-180" alt="intel-link-line-adviser" width="600px" />
 
 We need to insert the link for `BLAS_LIBS`, `LAPACK_LIBS`, and `SCALAPACK_LIBS`.
 We also need to find out where is the FFTW lib located. In NUS HPC, we can use

src/css/common.css

@@ -281,4 +281,7 @@ html[data-theme="dark"] {
   .footer {
     background-color: var(--ifm-background-surface-color);
   }
+  .inv-hue-rot-180 {
+    filter: invert(1) hue-rotate(180deg);
+  }
 }