Presentation: Added pdfpc zooming guide

wiki/android/episteme_reader.md

@@ -0,0 +1,13 @@
+# Episteme Reader
+
+[Episteme Reader](https://github.com/Aryan-Raj3112/episteme) is a free and open-source
+[e-book](/wiki/e-book.md) reader for [Android](/wiki/android.md).
+It supports annotations and OPDS catalogs and can handle large PDF files.
+For looking up words while reading, an offline dictionary application such as
+[OSS Dict](/wiki/android/oss_dict.md) can be used.
+Episteme Reader also features using translation apps like [DeepL](/wiki/deepl.md).
+
+## Setup
+
+On [Android](/wiki/android.md) Episteme Reader is available on [F-Droid](/wiki/android/f-droid.md)
+in a package called [Episteme Reader](https://f-droid.org/en/packages/com.aryan.reader.oss/).

wiki/android/f-droid.md

@@ -2,6 +2,8 @@
 
 [F-Droid](https://f-droid.org) is an installable catalogue of free and open source
 software for [Android](/wiki/android.md) systems.
+Alternatively, [Neo Store](/wiki/android/neo_store.md) can be used as a modern F-Droid client and
+alternative.
 
 ## Setup
 

wiki/android/neo_store.md

@@ -0,0 +1,10 @@
+# Neo Store
+
+[Neo Store](https://github.com/NeoApplications/Neo-Store) is a free and open-source
+[F-Droid](/wiki/android/f-droid.md) client and alternative for [Android](/wiki/android.md).
+It can be used to browse, install and update applications from F-Droid and other repositories.
+
+## Setup
+
+On [Android](/wiki/android.md) Neo Store is available on [F-Droid](/wiki/android/f-droid.md) in a
+package called [Neo Store](https://f-droid.org/packages/com.machiav3lli.fdroid/).

wiki/android/oss_dict.md

@@ -0,0 +1,22 @@
+# OSS Dict
+
+[OSS Dict](https://github.com/fengdeyingzi/OSS-Dict) is a free and open-source offline dictionary
+application for [Android](/wiki/android.md) and a fork of Aard 2.
+The application can for example be used together with
+[Episteme Reader](/wiki/android/episteme_reader.md) for looking up words while reading.
+
+## Setup
+
+On [Android](/wiki/android.md) OSS Dict is available on [F-Droid](/wiki/android/f-droid.md)
+in a package called [OSS-Dict](https://f-droid.org/en/packages/com.akylas.aard2/).
+
+## Usage 
+
+This section addresses the usage of OSS Dict.
+
+### Installing Dictionaries
+
+Dictionary files have to be downloaded separately and added to OSS Dict.
+This is usually done by using `.slob` files.
+A guide on where to find them can be found in the
+[wiki of the slob Github page](https://github.com/itkach/slob/wiki/Dictionaries).

wiki/calibre.md

@@ -154,9 +154,9 @@ This section focuses on the basic usage of calibre's web version.
 
 #### Problems Logging in With Reverse Proxy
 
-It is [a known problem](https://github.com/janeczku/calibre-web/issues/2916?utm_source=chatgpt.com)
+It is [a known problem](https://github.com/janeczku/calibre-web/issues/2916) not being able to login
-not being able to login when using a [reverse proxy](/wiki/reverse-proxy.md) like
+when using a [reverse proxy](/wiki/reverse-proxy.md) like [Nginx](/wiki/nginx.md) or
-[Nginx](/wiki/nginx.md) or [Traefik](/wiki/traefik.md).
+[Traefik](/wiki/traefik.md).
 In this case a 504 error will be displayed.
 According to various sources this can be fixed by trying previous sections until it works again.
 

wiki/deepl.md

@@ -0,0 +1,14 @@
+# DeepL
+
+[DeepL](https://www.deepl.com) is an AI-based translation service for translating text and documents
+between various languages.
+
+## Setup
+
+DeepL is a website that can be visited without additional setup needed.
+
+On [Android](/wiki/android.md) DeepL is available on [F-Droid](/wiki/android/f-droid.md) in a
+package called [DeepL](https://f-droid.org/en/packages/com.example.deeplviewer/).
+Additionally, it is available on the [Google Play store](/wiki/android.md#app-store) in a package
+called
+[DeepL Translate](https://play.google.com/store/apps/details?id=com.deepl.mobiletranslator&hl=en-US).

wiki/e-book.md

@@ -25,18 +25,18 @@ If the software is to be used by you check your local laws first.
 ## Tracking
 
 Many e-book readers include a feature to keep track of read books.
-However they do not track the time that was needed to finish them or the number of e-books read in 
+However they do not track the time that was needed to finish them or the number of e-books read in
 a specific timeframe.
 This can be done by using a special tracker like [Jelu](/wiki/jelu.md) which can be used to keep
 track of e-books and normal books at the same time which is a feature no e-book reader has.
 
-## E-Reader 
+## E-Reader
 
 An e-reader is a device that is designed for reading digital e-books.
 Some of them also feature [annotation tools](/wiki/note-taking.md).
 
-- [Episteme Reader](https://github.com/Aryan-Raj3112/episteme) is a native
+- [Episteme Reader](/wiki/android/episteme_reader.md) is a native [Android](/wiki/android.md)
-  [Android](/wiki/android.md) [e-book](/wiki/e-book.md) reader, that is capable of annotations.
+  e-book reader, that is capable of annotations.
   In contrast to [Saber](/wiki/note-taking.md#software) it is able to handle large PDF-files.
   Additionally it features the OPDS protocol for catalogs which can, for example, connect to
   [the OPDS catalog of calibre's web version](/wiki/calibre.md#connect-via-opds).

wiki/firefox.md

@@ -98,10 +98,10 @@ The following add-ons provide translation options.
   add-on made by Mozilla itself that gives an option for marked text to display its translations 
   quickly from the extension menu but can also translate whole sites.
 - [Simple Translate](https://addons.mozilla.org/en-US/firefox/addon/simple-translate) is an add-on
-  supporting both DeepL and Google Translate to translate marked text in various different ways that
+  supporting both [DeepL](/wiki/deepl.md) and Google Translate to translate marked text in various
-  can be automatically displayed.
+  different ways that can be automatically displayed.
 - [To DeepL translation](https://addons.mozilla.org/en-US/firefox/addon/to-deepl/) gives an option
-  for marked text to display its translation in a pop-up window using DeepL.
+  for marked text to display its translation in a pop-up window using [DeepL](/wiki/deepl.md).
 
 #### Shopping Add-ons
 

wiki/presentation.md

@@ -24,3 +24,7 @@ The following is a list of these tools.
     Basic guides for this can be found on the pages of
     [toromtomtom](https://www.toromtomtom.com/blog/2021/10/17/pdfpc.html) and
     [Utah university](https://ctan.math.utah.edu/ctan/tex-archive/macros/latex/contrib/pdfpc/pdfpc-doc.pdf).
+    Most importantly the tab key opens the overview mode, `p` pauses, `f` freezes, `t` opens 
+    clickable options, `g` opens the option to navigate to a specific page, `h` hides and `b` shows 
+    `a` blank foil. After selecting a rectangular space with the pointer tool, `z` can be pressed to
+    zoom into it.

wiki/programming_language/python.md

@@ -381,7 +381,7 @@ can be referenced.
 ### TensorFlow
 
 This section addresses the [TensorFlow module](https://www.tensorflow.org/).
-Tensorflos is a machine learning resource which is often used for
+TensorFlow is a machine learning resource which is often used for
 [neural networks](/wiki/neural_network.md).
 
 Apart from [package managers](/wiki/linux/package_manager.md) and
@@ -390,6 +390,88 @@ on [the official website](https://www.tensorflow.org/install/source).
 This may especially be useful if specific configurations are needed such as vendor specific GPU
 support.
 
+#### Setup TensorFlow with CUDA in a uv Project
+
+TensorFlow can install its required CUDA user-space libraries as optional dependencies on Linux.
+The NVIDIA driver still has to be installed on the host system.
+
+```sh
+nvidia-smi
+uv add 'tensorflow[and-cuda]'
+```
+
+Verify whether TensorFlow was built with CUDA support and detects the GPU.
+
+```sh
+uv run python -c 'import tensorflow as tf; print("CUDA build:", tf.test.is_built_with_cuda()); print(tf.config.list_physical_devices("GPU"))'
+```
+
+If `CUDA build` is `True` but no GPU is listed and TensorFlow reports that it cannot load GPU
+libraries, the dynamic linker may not find the NVIDIA libraries installed inside the virtual
+environment.
+The following command temporarily adds all library directories from the installed `nvidia-*`
+packages.
+
+```sh
+SITE_PACKAGES=$(uv run python -c 'import site; print(site.getsitepackages()[0])')
+CUDA_LIBS=$(find "$SITE_PACKAGES/nvidia" -type d -name lib -printf '%p:')
+
+LD_LIBRARY_PATH="${CUDA_LIBS}/usr/lib" uv run python -c \
+  'import tensorflow as tf; print(tf.config.list_physical_devices("GPU"))'
+```
+
+If the test succeeds, store the path in a machine-specific dotenv file.
+
+```sh
+printf 'LD_LIBRARY_PATH=%s/usr/lib\n' "$CUDA_LIBS" > .env.cuda
+printf '.env.cuda\n' >> .gitignore
+```
+
+Run the project with the file explicitly.
+
+```sh
+uv run --env-file .env.cuda python <script>.py
+```
+
+To load it automatically, create a `.envrc` file and use
+[direnv](https://direnv.net/).
+
+```sh
+printf 'export UV_ENV_FILE="$PWD/.env.cuda"\n' > .envrc
+eval "$(direnv hook zsh)"
+direnv allow
+```
+
+For more information about dotenv files in uv, refer to the
+[uv environment variable section](/wiki/programming_language/python/uv.md#loading-environment-variables).
+
+Some recent GPUs may require CUDA kernels to be compiled from PTX on the first run because the
+TensorFlow wheel does not yet contain native kernel binaries for their compute capability.
+For example, TensorFlow 2.21 reports this for an RTX 5060 Ti with compute capability `12.0a`.
+According to NVIDIA's
+[explanation of PTX compatibility](https://developer.nvidia.com/blog/understanding-ptx-the-assembly-language-of-cuda-gpu-computing/),
+embedded PTX can be compiled for newer GPU generations at runtime.
+The first start can therefore take considerably longer, while the resulting binary is normally
+cached for subsequent runs.
+
+Make sure the `ptxas` executable installed by the CUDA dependency is available in the virtual
+environment.
+
+```sh
+VENV_DIR=$(uv run python -c 'import sys; print(sys.prefix)')
+PTXAS=$(find "$VENV_DIR" -type f -name ptxas -print -quit)
+ln -sf "$PTXAS" "$VENV_DIR/bin/ptxas"
+```
+
+The library lookup problem is not caused by the GPU being new; it is an environment configuration
+issue.
+The new GPU generation only explains why TensorFlow falls back to PTX JIT compilation after the
+libraries have been found.
+
+Refer to the
+[official TensorFlow installation guide](https://www.tensorflow.org/install/pip) for the current
+CUDA installation and troubleshooting steps.
+
 #### Basic Usage of TensorFlow
 
 The basic usage of TensorFlow is described in

wiki/programming_language/python/uv.md

@@ -113,6 +113,44 @@ desired package.
 uv add <package>
 ```
 
+### Loading Environment Variables
+
+[`uv run`](https://docs.astral.sh/uv/concepts/configuration-files/#environment-variable-files) can
+load environment variables from dotenv files.
+Use `--env-file` to load a file for one command.
+
+```sh
+uv run --env-file .env.local python <script>.py
+```
+
+Alternatively, set `UV_ENV_FILE` to use the same file for subsequent `uv run` commands in the
+current shell.
+
+```sh
+export UV_ENV_FILE="$PWD/.env.local"
+uv run python <script>.py
+```
+
+For project-local automation, [direnv](https://direnv.net/) can set `UV_ENV_FILE` whenever the
+project directory is entered.
+Create `.envrc` in the project root with the following content.
+
+```sh
+export UV_ENV_FILE="$PWD/.env.local"
+```
+
+Enable the shell hook, then approve the file.
+The example uses Zsh.
+
+```sh
+eval "$(direnv hook zsh)"
+direnv allow
+```
+
+The shell hook should be added to `~/.zshrc` to enable it in new shells.
+`direnv allow` always expects a `.envrc`; it does not load `.env.local` directly.
+Machine-specific dotenv files should usually be added to `.gitignore`.
+
 ### Installing CLI Tools
 
 Besides managing projects and virtual environments, `uv` can also install

wiki/rembg.md

@@ -0,0 +1,45 @@
+# rembg
+
+[rembg](https://github.com/danielgatis/rembg) is a [command line](/wiki/linux/shell.md) tool and
+Python library for removing image backgrounds using pretrained
+[neural networks](/wiki/neural_network.md) executed through ONNX Runtime.
+
+## Setup
+
+rembg can be installed from source as described on the
+[official GitHub repository](https://github.com/danielgatis/rembg).
+
+Alternatively a [Python package manager](/wiki/programming_language/python.md#package-management)
+can be used to install rembg.
+For the backend one must select CPU or GPU usage.
+For command line usage an extra package has to be installed in addition to the inference backend.
+Depending on that choice the name of the package will be `"rembg[cpu,cli]"` or `rembg[gpu,cli]`.
+
+## Usage
+
+This section addresses the usage of rembg.
+
+### Remove Background
+
+The following command removes the background from an image using a specific `<model>`.
+In this case `<input>` is a placeholder for the input image path and `<output>` for the output image
+path.
+
+```sh
+rembg i -m <model> <input> <output>
+```
+
+Depending on the use-case other models may also be used.
+`u2net` is a general purpose model and most of the time a good default.
+`u2net_human_seg` is optimized for human subjects.
+`isnet-general-use` is an alternative general-purpose model.
+
+The quality depends on the selected model and the input image.
+For portraits with complex backgrounds such as bushes or trees, `u2net` often produces the best
+overall results, while other models may perform better on different types of images.
+
+Additionally, alpha matting can improve difficult edges such as hair.
+
+```sh
+rembg i -m <model> -a <input> <output>
+```