Restore old docs and workflow (#140)

# Description

There's some problems with the new version of docs (listed in the linked
issue). As such, for now `pytket-cutensornet` will continue maintaining
the old version of the docs, and using it as the main version. Once
these issues are solved, I'm happy to make another release switching to
the new docs.

# Related issues

#139

.github/workflows/build_and_test.yml

@@ -132,3 +132,23 @@ jobs:
         cd .github/workflows/docs
         mkdir extensions
         ./build-docs -d ${GITHUB_WORKSPACE}/.github/workflows/docs/extensions/api
+    - name: Upload docs as artefact
+      uses: actions/upload-pages-artifact@v3
+      with:
+        path: .github/workflows/docs/extensions
+
+  publish_docs:
+    name: Publish docs
+    if: github.event_name == 'release'
+    needs: build_docs
+    runs-on: ubuntu-22.04
+    permissions:
+      pages: write
+      id-token: write
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+    steps:
+    - name: Deploy to GitHub Pages
+      id: deployment
+      uses: actions/[email protected]

README.md

@@ -24,7 +24,7 @@ Currently, only single-GPU calculations are supported, but a multi-GPU execution
 implemented in the due course using `mpi4py` library.
 
 Some useful links:
-- [API Documentation](https://tket.quantinuum.com/extensions/pytket-cutensornet/)
+- [API Documentation](https://cqcl.github.io/pytket-cutensornet/api/index.html)
 
 ## Getting started
 

docs/changelog.rst

@@ -4,7 +4,6 @@ Changelog
 0.7.0 (July 2024)
 -----------------
 
-* Updated pytket version requirement to 1.30.
 * API breaking changes
     * Renamed ``CuTensorNetBackend`` to ``CuTensorNetStateBackend``.
     * Moved ``get_operator_expectation_value`` and ``get_circuit_overlap`` from ``backends`` submodule to ``general_state`` submodule.
@@ -18,6 +17,7 @@ Changelog
 * New feature: a seed can now be provided to ``Config`` objects, providing reproducibility across ``StructuredState`` simulations.
 * New feature: ``apply_unitary`` both for ``MPS`` and ``TTN`` to apply an arbitrary unitary matrix, rather than a ``pytket.Command``.
 * New feature: ``apply_qubit_relabelling`` both for ``MPS`` and ``TTN`` to change the name of their qubits. This is now used within ``simulate`` to take into account the action of implicit SWAPs in pytket circuits (no additional SWAP gates are applied).
+* Updated pytket version requirement to 1.30.
 
 0.6.1 (April 2024)
 ------------------