ci: Allow specifying hugo version

Because of
hugo: /usr/lib/x86_64-linux-gnu/libstdc++.so.6: version `GLIBCXX_3.4.29' not found (required by hugo)

.forgejo/workflows/deploy.yml

@@ -0,0 +1,60 @@
+name: Deploy Hugo Site
+
+on:
+  push:
+    branches:
+      - main
+      - forge-ci
+
+jobs:
+  build-and-deploy:
+    runs-on: ubuntu-24.04
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          submodules: recursive
+          fetch-depth: 0
+
+      - name: Install Hugo - latest version
+        run: |
+          apt-get update -y
+          apt-get install -y wget tar jq # Install necessary components
+          ldd --version
+          
+          # If a Hugo version is provided in the secrets, use this
+          # else latest will be used
+          if [ -n "${{ secrets.HUGO_VERSION }}" ]; then
+            HUGO_VERSION="${{ secrets.HUGO_VERSION }}"
+            echo "Using Hugo version from secret: $HUGO_VERSION"
+          else
+          # Use the GitHub API to get information about the latest release, then use jq to find out the tag name
+            HUGO_VERSION=$(wget -qO- https://api.github.com/repos/gohugoio/hugo/releases/latest | jq -r '.tag_name')
+            echo "Using latest Hugo version: $HUGO_VERSION"
+          fi
+          
+          # Use ${HUGO_VERSION#v} to strip the v from v1.0.0
+          # See "Substring Removal" in https://tldp.org/LDP/abs/html/string-manipulation.html
+          wget -O hugo.tar.gz "https://github.com/gohugoio/hugo/releases/download/${HUGO_VERSION}/hugo_extended_${HUGO_VERSION#v}_Linux-64bit.tar.gz"
+          tar -xzf hugo.tar.gz hugo
+          mv hugo /usr/local/bin/hugo
+          chmod +x /usr/local/bin/hugo
+          ldd /usr/local/bin/hugo
+          hugo version
+
+      - name: Build site
+        run: hugo --minify
+
+      - name: Deploy to server via rsync
+        env:
+          DEPLOY_HOST: ${{ secrets.DEPLOY_HOST }}
+          DEPLOY_USER: ${{ secrets.DEPLOY_USER }}
+          DEPLOY_PATH: ${{ secrets.DEPLOY_PATH }}
+        run: |
+          apt-get install -y rsync openssh-client
+          mkdir -p ~/.ssh
+          echo "${{ secrets.SSH_PRIVATE_KEY }}" > ~/.ssh/id_ed25519
+          chmod 600 ~/.ssh/id_ed25519
+          ssh-keyscan -H "$DEPLOY_HOST" >> ~/.ssh/known_hosts
+          rsync -avz --delete public/ "$DEPLOY_USER@$DEPLOY_HOST:$DEPLOY_PATH"

config.toml

@@ -7,7 +7,6 @@ disqusShortname = ""
 # Enable Google Analytics by entering your tracking code
 googleAnalytics = ""
 preserveTaxonomyNames = true
-paginate = 5 #frontpage pagination
 
 [privacy]
   # Google Analytics privacy settings - https://gohugo.io/about/hugo-and-gdpr/index.html#googleanalytics

content/post/checking-shelters/index.md

@@ -0,0 +1,122 @@
+---
+title: "How to manually check hundreds of animal shelters - every 14 days"
+date: 2025-11-08T12:05:10+02:00
+lastmod: 2025-11-08T21:05:10+02:00
+draft: false
+image: "uploads/checking-shelters.png"
+categories: [ 'English' ]
+tags: [ 'notfellchen', 'animal shelter', 'animal welfare', 'django' ]
+---
+
+I run a website called [Notfellchen](https://notfellchen.org) that list animals that are waiting for adoption. It's
+currently restricted to fancy rats in Germany and that for good reason: Running this website involves **checking every
+shelter every two weeks manually**. You need to visit the website, check if there are new animals, contact the shelter
+and add them to notfellchen if they allow it. This takes time. A lot.
+
+This blog post will outline some of the things I did in order to streamline this and make it possible to **check every
+german shelter in 2.5 hours**.
+
+## General process
+
+When you establish a process. want others to help you or if you want to find inefficiencies, it's a good idea to
+formalize it. So here is a rough BPMN diagram of the whole process.
+
+{{< html animal-discovery.drawio.html >}}
+
+## List of animal shelters
+
+Focusing on the first step: We want to check the website of an animal shelter - but where do we get a list of animal
+shelters from? Luckily there is an easy answer: [OpenStreetMap](https://openstreetmap.org) and I wrote a
+whole [other blog post on how I imported and improved this data](https://hyteck.de/post/improve-osm-by-using-it/).
+
+## Species-specific link
+
+Importing this data provides us (most of the time) with a link to the shelter's website. However, rats are usually not
+listed on the home page but on a subsite.
+In order to save time, I introduced the concept of a species-specific link per organization and species.
+
+So for the Tierheim Entenhausen this might look like this
+
+| Species | Species specific link                                  |
+|---------|--------------------------------------------------------| 
+| Cat     | https://tierheim-entenhausen.de/adoption/cats          |
+| Rats    | https://tierheim-entenhausen.de/adoption/small-mammals |
+
+As animal shelter pages look very different from each other, clicking this link provides an enormous time benefit
+compared to clicking through a homepage manually.
+
+# Org check page
+
+I set up a special page to make it most efficient to check shelters. It's structured in four parts:
+
+* **Stats**: The stats show how many animal shelters are checked in the last two weeks and how many to go.
+* **Not checked for the longest period**: Shows the animal shelters to check next, it's therefore sorted by the date
+  they were last checked
+* **In active communication**: A overview of the organizations where there is communication (or an attempt thereof).
+  This can take multiple das or even weeks so the internal comment field is very useful to keep track.
+* **Last checked** It sometimes happens that I accidentally set a organization to "Checked" by accident. I added this
+  section to make it easier to revert that.
+
+![](screenshot-checking-site.png)
+
+## Shortcuts
+
+To make it even faster to work through the organizations I added some shortcuts for the most common functionality and
+documented the browser own shortcut to close a tab.
+
+* `O`: Open website of the first organization
+* `CTRL+W`: Close tab (Firefox, Chrome)
+* `C`: Mark first organization as checked
+
+## Results
+
+After implementing all this, how long does it take now to check all organizations? Here are the numbers
+
+| Measurement                                               |              |
+|-----------------------------------------------------------|--------------|
+| Time to check one organization (avg.)                     | 12.1s        |
+| Organization checked per minute                           | 4.96 org/min |
+| Time to check all (eligible) german animal shelters (429) | 1 h 16 min   | 
+
+This excludes the time, it takes to add animals or contact rescue organizations. One of these actions must be taken
+whenever an eligible animal is found on a website. Here you can see how this interrupts the process:
+
+![](progress.png)
+
+And here is the breakdown of time per activity. A big caveat here is, that I did not follow up on previous conversations
+here, therefore the contacting number is likely an underestimation.
+
+| Activity   | Time spent | Percentage |
+|------------|------------|------------|
+| Checking   | 54 min 44s | 72.3%      |
+| Adding     | 11 min 15s | 14.9%      |
+| Contacting | 9min 41s   | 12.8%      |
+
+To me, this looks like a pretty good result. I can't say which optimizations brought how much improvement, but I'd argue
+they all play a role in reaching the 12s per rescue organizations that is checked.
+
+In order to check all german animal shelters, one needs to put in about 2 and a half hours every two weeks. That seems
+reasonable to me. Further improvements of the likely do not lie in the organization check page but the contact process
+and adoption notice form.
+
+For now, I'm happy with the results.
+
+# Addendum: Common annoyances
+
+When doing this over the last few months I encountered some recurring issues that not only were annoying but also take
+up a majority of the time. Here are some that stood out
+
+* **Broken SSL encryption** So many animal shelters do not have a functioning SSL certificate. It takes time to work
+  around the warnings.
+* **No results not indicated** More often than not, animal shelters do not have rats. However, when you visit a page
+  like [this](https://tierschutzliga.de/tierheime/tierparadies-oberdinger-moos/tiervermittlung/#?THM=TOM&Tierart=Kleintiere)
+  it's hard to know if there is a technical issue or if there are no animals for your search.
+* **No static links** Sites where you have to click through a menu to get to the right page, but you can not link
+  directly to it.
+* **No website** Seriously, there are some animal shelters that only use Instagram or Facebook to tell people about the
+  animals they have. This is not only a privacy nightmare, it's also incredibly hard to find out which information is
+  up-to-date. Furthermore, there exists no data structure, so posts about animals often miss crucial information like
+  the sex.
+
+While I obviously have some grievances here, I know the organizations never have enough resources, and they'd
+usually love to have a nicer website. Just keep that in mind too.

layouts/shortcodes/html.html

@@ -0,0 +1,14 @@
+{{ $source := index .Params 0 }}
+<div class="embedded-html">
+    {{ with .Page.Resources.GetMatch $source | readFile }}
+    {{ replace . "https://viewer.diagrams.net/js/viewer-static.min.js" "/js/viewer-static.min.js"|safeHTML }}
+    {{ end }}
+</div>
+<style>
+    /* Reset the image width that is set in the theme for
+    If wired styling issues appear, check if the theme is responsible and eventually unset this here
+    */
+    .geAdaptiveAsset {
+        width: unset;
+    }
+</style>