Various fixes for issues with new architecture (#208)

* Add formatters to the TOC - fixes #204

* Add 'steps' settings to the example YAML in the docs. Fixes #206

* Improved docs on authentication architecture

* Fix setting modules on the command line - they now override any module settings in the orchestration as opposed to appending

* Fix tests for gsheet-feeder: add a test service_account.json (note: not real keys in there)

* Rename the command line entrypoint to _command_line_run

Also: make it clear that code implementation should not call this
Make sure the command line entry returns (we don't want a generator)

* Fix unit tests to use now code-entry points

* Version bump

* Move iterating of generator up to __main__

* Breakpoint

* two minor fixes

* Fix unit tests + add new '__main__' entry point implementation test

* Skip youtube tests if running on CI. Should still run them locally

* Fix full implementation run on GH actions

* Fix skipif test for GH Actions CI

* Add skipifs for truth - it blocks GH:

---------

Co-authored-by: msramalho <19508417+msramalho@users.noreply.github.com>

src/auto_archiver/core/base_module.py

@@ -63,12 +63,6 @@ class BaseModule(ABC):
     def config_setup(self, config: dict):
 
         authentication = config.get('authentication', {})
-        # extract out concatenated sites
-        for key, val in copy(authentication).items():
-            if "," in key:
-                for site in key.split(","):
-                    authentication[site] = val
-                del authentication[key]
 
         # this is important. Each instance is given its own deepcopied config, so modules cannot
         # change values to affect other modules
@@ -89,16 +83,21 @@ class BaseModule(ABC):
         Returns the authentication information for a given site. This is used to authenticate
         with a site before extracting data. The site should be the domain of the site, e.g. 'twitter.com'
         
-        extract_cookies: bool - whether or not to extract cookies from the given browser and return the 
-        cookie jar (disabling can speed up) processing if you don't actually need the cookies jar
+        :param site: the domain of the site to get authentication information for
+        :param extract_cookies: whether or not to extract cookies from the given browser/file and return the cookie jar (disabling can speed up processing if you don't actually need the cookies jar).
 
-        Currently, the dict can have keys of the following types:
-        - username: str - the username to use for login
-        - password: str - the password to use for login
-        - api_key: str - the API key to use for login
-        - api_secret: str - the API secret to use for login
-        - cookie: str - a cookie string to use for login (specific to this site)
-        - cookies_jar: YoutubeDLCookieJar | http.cookiejar.MozillaCookieJar - a cookie jar compatible with requests (e.g. `requests.get(cookies=cookie_jar)`)
+        :returns: authdict dict of login information for the given site
+
+        **Global options:**\n
+        * cookies_from_browser: str - the name of the browser to extract cookies from (e.g. 'chrome', 'firefox' - uses ytdlp under the hood to extract\n
+        * cookies_file: str - the path to a cookies file to use for login\n
+
+        **Currently, the sites dict can have keys of the following types:**\n
+        * username: str - the username to use for login\n
+        * password: str - the password to use for login\n
+        * api_key: str - the API key to use for login\n
+        * api_secret: str - the API secret to use for login\n
+        * cookie: str - a cookie string to use for login (specific to this site)\n
         """
         # TODO: think about if/how we can deal with sites that have multiple domains (main one is x.com/twitter.com)
         # for now the user must enter them both, like "x.com,twitter.com" in their config. Maybe we just hard-code?