Revert "remove incorrect end-to-end implementation"

This reverts commit ebc5e31 from instructlab#76.

Here is my comment on instructlab#76 that better explains how this workflow fits
in to the bigger picture:

> Testing the user end-to-end instructlab workflow is the point of
> this workflow. It will give you an early indication if your changes to
> the library break the CLI. Sometimes that might be on purpose, but in
> case it's not, it will let you know before you publish a release.
>
> This is not intended to replace other functional testing aimed at the
> library more directly, but the full end-to-end workflow is important.
>
> I'm OK if you want to disable it for the moment, but it needs to go
> back on at some point. A better way to do that is to just edit the
> workflow to turn it off instead of removing it from the repo.

Issue instructlab#63

Signed-off-by: Russell Bryant <rbryant@redhat.com>

.github/workflows/e2e.yml

@@ -0,0 +1,92 @@
+# SPDX-License-Identifier: Apache-2.0
+
+name: E2E test
+
+on:
+  push:
+    branches:
+      - "main"
+      - "release-**"
+    paths:
+      - '**.py'
+      - 'pyproject.toml'
+      - 'requirements*.txt'
+      - '.github/workflows/e2e.yml'
+  pull_request:
+    branches:
+      - "main"
+      - "release-**"
+    paths:
+      - '**.py'
+      - 'pyproject.toml'
+      - 'requirements*.txt'
+      - '.github/workflows/e2e.yml'
+
+jobs:
+  e2e:
+    runs-on: ubuntu-gpu
+
+    permissions:
+      pull-requests: write
+
+    steps:
+      # No step-security/harden-runner since this is a self-hosted runner
+      - name: Checkout
+        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332 # v4.1.7
+        with:
+          # https://github.com/actions/checkout/issues/249
+          fetch-depth: 0
+
+      - name: Install Packages
+        run: |
+          sudo apt-get install -y cuda-toolkit git cmake build-essential virtualenv
+          nvidia-smi
+          ls -l /dev/nvidia*
+
+      - name: Setup Python 3.11
+        uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d # v5.1.0
+        with:
+          python-version: 3.11
+          cache: pip
+          cache-dependency-path: |
+            **/pyproject.toml
+            **/requirements*.txt
+
+      - name: Remove llama-cpp-python from cache
+        run: |
+          pip cache remove llama_cpp_python
+
+      - name: Cache huggingface
+        uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 # v4.0.2
+        with:
+          path: ~/.cache/huggingface
+          # config contains DEFAULT_MODEL
+          key: huggingface-${{ hashFiles('src/instructlab/configuration.py') }}
+
+      - name: Install instructlab and libraries
+        run: |
+          export PATH="/home/runner/.local/bin:/usr/local/cuda/bin:$PATH"
+          python3 -m venv venv
+          . venv/bin/activate
+          git clone https://github.com/instructlab/instructlab.git
+          cd instructlab
+          sed 's/\[.*\]//' requirements.txt > constraints.txt
+          python3 -m pip cache remove llama_cpp_python
+          CMAKE_ARGS="-DLLAMA_CUBLAS=on" python3 -m pip install --no-binary llama_cpp_python -c constraints.txt llama_cpp_python
+          # needed for --4-bit-quant option to ilab train
+          python3 -m pip install bitsandbytes
+          # install instructlab
+          python3 -m pip install .
+          cd ..
+          # Install instructlab-sdg
+          python3 -m pip install .
+
+      - name: Run e2e test
+        run: |
+          . venv/bin/activate
+          ./instructlab/scripts/basic-workflow-tests.sh -cm
+
+      - name: Remove llama-cpp-python from cache
+        if: always()
+        run: |
+          pip cache remove llama_cpp_python

.github/workflows/spellcheck.yml

@@ -0,0 +1,45 @@
+# SPDX-License-Identifier: Apache-2.0
+
+name: Spellcheck
+
+on:
+  push:
+    branches:
+      - "main"
+    paths:
+      - '**.md'
+      - '.github/workflows/spellcheck.yml' # This workflow
+  pull_request:
+    branches:
+      - "main"
+    paths:
+      - '**.md'
+      - '.github/workflows/spellcheck.yml' # This workflow
+
+env:
+  LC_ALL: en_US.UTF-8
+
+defaults:
+  run:
+    shell: bash
+
+permissions:
+  contents: read
+
+jobs:
+  spellcheck:
+    name: Spellcheck (en_US)
+    runs-on: ubuntu-latest
+    steps:
+      - name: "Harden Runner"
+        uses: step-security/harden-runner@17d0e2bd7d51742c71671bd19fa12bdc9d40a3d6 # v2.8.1
+        with:
+         egress-policy: audit # TODO: change to 'egress-policy: block' after couple of runs
+
+      - name: "Checkout"
+        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332 # v4.1.7
+        with:
+          fetch-depth: 0
+
+      - name: Spellcheck
+        uses: rojopolis/spellcheck-github-actions@d354a4dc525c8067555c7481b60416cedb0060ff # v0.38.0