Merge remote-tracking branch 'origin/main' into heroku-app

README.md

@@ -15,11 +15,12 @@ Contents
 3. [Install/Deploy](#install)
     1. [Heroku Quick Deploy](#a-heroku-quick-deploy)
     2. [Repl.it](#b-replit)
-    3. [pipx](#c-pipx)
-    4. [pip](#d-pip)
-    5. [Manual](#e-manual)
-    6. [Docker](#f-manual-docker)
-    7. [Arch/AUR](#arch-linux--arch-based-distributions)
+    3. [Fly.io](#c-flyio)
+    4. [pipx](#d-pipx)
+    5. [pip](#e-pip)
+    6. [Manual](#f-manual)
+    7. [Docker](#g-manual-docker)
+    8. [Arch/AUR](#arch-linux--arch-based-distributions)
 4. [Environment Variables and Configuration](#environment-variables)
 5. [Usage](#usage)
 6. [Extra Steps](#extra-steps)
@@ -90,7 +91,33 @@ Provides:
     - Supports custom domains
 - Downtime after periods of inactivity \([solution 1](https://repl.it/talk/ask/use-this-pingmat1replco-just-enter/28821/101298), [solution 2](https://repl.it/talk/learn/How-to-use-and-setup-UptimeRobot/9003)\)
 
-### C) [pipx](https://github.com/pipxproject/pipx#install-pipx)
+### C) [Fly.io](https://fly.io)
+
+You will need a [Fly.io](https://fly.io) account to do this. Fly requires a credit card to deploy anything, but you can have up to 3 shared-CPU VMs running full-time each month for free.
+
+#### Install the CLI:
+
+```bash
+curl -L https://fly.io/install.sh | sh
+```
+
+#### Deploy your app
+
+```bash
+fly apps create --org personal --port 5000
+# Choose a name and the Image builder
+# Enter `benbusby/whoogle-search:latest` as the image name
+fly deploy
+```
+
+Your app is now available at `https://<app-name>.fly.dev`.
+
+You can customize the `fly.toml`:
+- Remove the non-https service
+- Add environment variables under the `[env]` key
+  - Use `fly secrets set NAME=value` for more sensitive values like `WHOOGLE_PASS` and `WHOOGLE_PROXY_PASS`.
+
+### D) [pipx](https://github.com/pipxproject/pipx#install-pipx)
 Persistent install:
 
 `pipx install git+https://github.com/benbusby/whoogle-search.git`
@@ -99,7 +126,7 @@ Sandboxed temporary instance:
 
 `pipx run --spec git+https://github.com/benbusby/whoogle-search.git whoogle-search`
 
-### D) pip
+### E) pip
 `pip install whoogle-search`
 
 ```bash
@@ -126,7 +153,7 @@ optional arguments:
 ```
 See the [available environment variables](#environment-variables) for additional configuration.
 
-### E) Manual
+### F) Manual
 Clone the repo and run the following commands to start the app in a local-only environment:
 
 ```bash
@@ -183,7 +210,7 @@ sudo systemctl enable whoogle
 sudo systemctl start whoogle
 ```
 
-### F) Manual (Docker)
+### G) Manual (Docker)
 1. Ensure the Docker daemon is running, and is accessible by your user account
   - To add user permissions, you can execute `sudo usermod -aG docker yourusername`
   - Running `docker ps` should return something besides an error. If you encounter an error saying the daemon isn't running, try `sudo systemctl start docker` (Linux) or ensure the docker tool is running (Windows/macOS).

app/__init__.py

@@ -22,7 +22,7 @@ app.default_key = generate_user_key()
 app.no_cookie_ips = []
 app.config['SECRET_KEY'] = os.urandom(32)
 app.config['SESSION_TYPE'] = 'filesystem'
-app.config['VERSION_NUMBER'] = '0.5.1'
+app.config['VERSION_NUMBER'] = '0.5.2'
 app.config['APP_ROOT'] = os.getenv(
     'APP_ROOT',
     os.path.dirname(os.path.abspath(__file__)))

app/filter.py

@@ -22,6 +22,21 @@ def strip_blocked_sites(query: str) -> str:
     return query[:query.find('-site:')] if '-site:' in query else query
 
 
+def extract_q(q_str: str, href: str) -> str:
+    """Extracts the 'q' element from a result link. This is typically
+    either the link to a result's website, or a string.
+
+    Args:
+        q_str: The result link to parse
+        href: The full url to check for standalone 'q' elements first,
+              rather than parsing the whole query string and then checking.
+
+    Returns:
+        str: The 'q' element of the link, or an empty string
+    """
+    return parse_qs(q_str)['q'][0] if ('&q=' in href or '?q=' in href) else ''
+
+
 class Filter:
     def __init__(self, user_key: str, mobile=False, config=None) -> None:
         if config is None:
@@ -223,20 +238,18 @@ class Filter:
             link['target'] = '_blank'
 
         result_link = urlparse.urlparse(href)
-        query = parse_qs(
-            result_link.query
-        )['q'][0] if 'q=' in href else ''
+        q = extract_q(result_link.query, href)
 
-        if query.startswith('/'):
+        if q.startswith('/'):
             # Internal google links (i.e. mail, maps, etc) should still
             # be forwarded to Google
-            link['href'] = 'https://google.com' + query
+            link['href'] = 'https://google.com' + q
         elif '/search?q=' in href:
             # "li:1" implies the query should be interpreted verbatim,
             # which is accomplished by wrapping the query in double quotes
             if 'li:1' in href:
-                query = '"' + query + '"'
-            new_search = 'search?q=' + self.encrypt_path(query)
+                q = '"' + q + '"'
+            new_search = 'search?q=' + self.encrypt_path(q)
 
             query_params = parse_qs(urlparse.urlparse(href).query)
             for param in VALID_PARAMS:
@@ -247,7 +260,7 @@ class Filter:
             link['href'] = new_search
         elif 'url?q=' in href:
             # Strip unneeded arguments
-            link['href'] = filter_link_args(query)
+            link['href'] = filter_link_args(q)
 
             # Add no-js option
             if self.nojs:
@@ -255,7 +268,7 @@ class Filter:
         else:
             if href.startswith(MAPS_URL):
                 # Maps links don't work if a site filter is applied
-                link['href'] = MAPS_URL + "?q=" + strip_blocked_sites(query)
+                link['href'] = MAPS_URL + "?q=" + strip_blocked_sites(q)
             else:
                 link['href'] = href
 

setup.py

@@ -8,7 +8,7 @@ setuptools.setup(
     author='Ben Busby',
     author_email='benbusby@protonmail.com',
     name='whoogle-search',
-    version='0.5.1',
+    version='0.5.2',
     include_package_data=True,
     install_requires=requirements,
     description='Self-hosted, ad-free, privacy-respecting metasearch engine',