+.SS youtube\-dl coding conventions
+.PP
+This section introduces a guide lines for writing idiomatic, robust and
+future\-proof extractor code.
+.PP
+Extractors are very fragile by nature since they depend on the layout of
+the source data provided by 3rd party media hosters out of your control
+and this layout tends to change.
+As an extractor implementer your task is not only to write code that
+will extract media links and metadata correctly but also to minimize
+dependency on the source\[aq]s layout and even to make the code foresee
+potential future changes and be ready for that.
+This is important because it will allow the extractor not to break on
+minor layout changes thus keeping old youtube\-dl versions working.
+Even though this breakage issue is easily fixed by emitting a new
+version of youtube\-dl with a fix incorporated, all the previous
+versions become broken in all repositories and distros\[aq] packages
+that may not be so prompt in fetching the update from us.
+Needless to say, some non rolling release distros may never receive an
+update at all.
+.SS Mandatory and optional metafields
+.PP
+For extraction to work youtube\-dl relies on metadata your extractor
+extracts and provides to youtube\-dl expressed by an information
+dictionary (https://github.com/rg3/youtube-dl/blob/7f41a598b3fba1bcab2817de64a08941200aa3c8/youtube_dl/extractor/common.py#L94-L303)
+or simply \f[I]info dict\f[].
+Only the following meta fields in the \f[I]info dict\f[] are considered
+mandatory for a successful extraction process by youtube\-dl:
+.IP \[bu] 2
+\f[C]id\f[] (media identifier)
+.IP \[bu] 2
+\f[C]title\f[] (media title)
+.IP \[bu] 2
+\f[C]url\f[] (media download URL) or \f[C]formats\f[]
+.PP
+In fact only the last option is technically mandatory (i.e.
+if you can\[aq]t figure out the download location of the media the
+extraction does not make any sense).
+But by convention youtube\-dl also treats \f[C]id\f[] and \f[C]title\f[]
+as mandatory.
+Thus the aforementioned metafields are the critical data that the
+extraction does not make any sense without and if any of them fail to be
+extracted then the extractor is considered completely broken.
+.PP
+Any
+field (https://github.com/rg3/youtube-dl/blob/7f41a598b3fba1bcab2817de64a08941200aa3c8/youtube_dl/extractor/common.py#L188-L303)
+apart from the aforementioned ones are considered \f[B]optional\f[].
+That means that extraction should be \f[B]tolerant\f[] to situations
+when sources for these fields can potentially be unavailable (even if
+they are always available at the moment) and \f[B]future\-proof\f[] in
+order not to break the extraction of general purpose mandatory fields.
+.SS Example
+.PP
+Say you have some source dictionary \f[C]meta\f[] that you\[aq]ve
+fetched as JSON with HTTP request and it has a key \f[C]summary\f[]:
+.IP
+.nf
+\f[C]
+meta\ =\ self._download_json(url,\ video_id)
+\f[]
+.fi
+.PP
+Assume at this point \f[C]meta\f[]\[aq]s layout is:
+.IP
+.nf
+\f[C]
+{
+\ \ \ \ ...
+\ \ \ \ "summary":\ "some\ fancy\ summary\ text",
+\ \ \ \ ...
+}
+\f[]
+.fi
+.PP
+Assume you want to extract \f[C]summary\f[] and put it into the
+resulting info dict as \f[C]description\f[].
+Since \f[C]description\f[] is an optional meta field you should be ready
+that this key may be missing from the \f[C]meta\f[] dict, so that you
+should extract it like:
+.IP
+.nf
+\f[C]
+description\ =\ meta.get(\[aq]summary\[aq])\ \ #\ correct
+\f[]
+.fi
+.PP
+and not like:
+.IP
+.nf
+\f[C]
+description\ =\ meta[\[aq]summary\[aq]]\ \ #\ incorrect
+\f[]
+.fi
+.PP
+The latter will break extraction process with \f[C]KeyError\f[] if
+\f[C]summary\f[] disappears from \f[C]meta\f[] at some later time but
+with the former approach extraction will just go ahead with
+\f[C]description\f[] set to \f[C]None\f[] which is perfectly fine
+(remember \f[C]None\f[] is equivalent to the absence of data).
+.PP
+Similarly, you should pass \f[C]fatal=False\f[] when extracting optional
+data from a webpage with \f[C]_search_regex\f[],
+\f[C]_html_search_regex\f[] or similar methods, for instance:
+.IP
+.nf
+\f[C]
+description\ =\ self._search_regex(
+\ \ \ \ r\[aq]<span[^>]+id="title"[^>]*>([^<]+)<\[aq],
+\ \ \ \ webpage,\ \[aq]description\[aq],\ fatal=False)
+\f[]
+.fi
+.PP
+With \f[C]fatal\f[] set to \f[C]False\f[] if \f[C]_search_regex\f[]
+fails to extract \f[C]description\f[] it will emit a warning and
+continue extraction.
+.PP
+You can also pass \f[C]default=<some\ fallback\ value>\f[], for example:
+.IP
+.nf
+\f[C]
+description\ =\ self._search_regex(
+\ \ \ \ r\[aq]<span[^>]+id="title"[^>]*>([^<]+)<\[aq],
+\ \ \ \ webpage,\ \[aq]description\[aq],\ default=None)
+\f[]
+.fi
+.PP
+On failure this code will silently continue the extraction with
+\f[C]description\f[] set to \f[C]None\f[].
+That is useful for metafields that may or may not be present.
+.SS Provide fallbacks
+.PP
+When extracting metadata try to do so from multiple sources.
+For example if \f[C]title\f[] is present in several places, try
+extracting from at least some of them.
+This makes it more future\-proof in case some of the sources become
+unavailable.
+.SS Example
+.PP
+Say \f[C]meta\f[] from the previous example has a \f[C]title\f[] and you
+are about to extract it.
+Since \f[C]title\f[] is a mandatory meta field you should end up with
+something like:
+.IP
+.nf
+\f[C]
+title\ =\ meta[\[aq]title\[aq]]
+\f[]
+.fi
+.PP
+If \f[C]title\f[] disappears from \f[C]meta\f[] in future due to some
+changes on the hoster\[aq]s side the extraction would fail since
+\f[C]title\f[] is mandatory.
+That\[aq]s expected.
+.PP
+Assume that you have some another source you can extract \f[C]title\f[]
+from, for example \f[C]og:title\f[] HTML meta of a \f[C]webpage\f[].
+In this case you can provide a fallback scenario:
+.IP
+.nf
+\f[C]
+title\ =\ meta.get(\[aq]title\[aq])\ or\ self._og_search_title(webpage)
+\f[]
+.fi
+.PP
+This code will try to extract from \f[C]meta\f[] first and if it fails
+it will try extracting \f[C]og:title\f[] from a \f[C]webpage\f[].
+.SS Regular expressions
+.SS Don\[aq]t capture groups you don\[aq]t use
+.PP
+Capturing group must be an indication that it\[aq]s used somewhere in
+the code.
+Any group that is not used must be non capturing.
+.SS Example
+.PP
+Don\[aq]t capture id attribute name here since you can\[aq]t use it for
+anything anyway.
+.PP
+Correct:
+.IP
+.nf
+\f[C]
+r\[aq](?:id|ID)=(?P<id>\\d+)\[aq]
+\f[]
+.fi
+.PP
+Incorrect:
+.IP
+.nf
+\f[C]
+r\[aq](id|ID)=(?P<id>\\d+)\[aq]
+\f[]
+.fi
+.SS Make regular expressions relaxed and flexible
+.PP
+When using regular expressions try to write them fuzzy, relaxed and
+flexible, skipping insignificant parts that are more likely to change,
+allowing both single and double quotes for quoted values and so on.
+.SS Example
+.PP
+Say you need to extract \f[C]title\f[] from the following HTML code:
+.IP
+.nf
+\f[C]
+<span\ style="position:\ absolute;\ left:\ 910px;\ width:\ 90px;\ float:\ right;\ z\-index:\ 9999;"\ class="title">some\ fancy\ title</span>
+\f[]
+.fi
+.PP
+The code for that task should look similar to:
+.IP
+.nf
+\f[C]
+title\ =\ self._search_regex(
+\ \ \ \ r\[aq]<span[^>]+class="title"[^>]*>([^<]+)\[aq],\ webpage,\ \[aq]title\[aq])
+\f[]
+.fi
+.PP
+Or even better:
+.IP
+.nf
+\f[C]
+title\ =\ self._search_regex(
+\ \ \ \ r\[aq]<span[^>]+class=(["\\\[aq]])title\\1[^>]*>(?P<title>[^<]+)\[aq],
+\ \ \ \ webpage,\ \[aq]title\[aq],\ group=\[aq]title\[aq])
+\f[]
+.fi
+.PP
+Note how you tolerate potential changes in the \f[C]style\f[]
+attribute\[aq]s value or switch from using double quotes to single for
+\f[C]class\f[] attribute:
+.PP
+The code definitely should not look like:
+.IP
+.nf
+\f[C]
+title\ =\ self._search_regex(
+\ \ \ \ r\[aq]<span\ style="position:\ absolute;\ left:\ 910px;\ width:\ 90px;\ float:\ right;\ z\-index:\ 9999;"\ class="title">(.*?)</span>\[aq],
+\ \ \ \ webpage,\ \[aq]title\[aq],\ group=\[aq]title\[aq])
+\f[]
+.fi
+.SS Long lines policy
+.PP
+There is a soft limit to keep lines of code under 80 characters long.
+This means it should be respected if possible and if it does not make
+readability and code maintenance worse.
+.PP
+For example, you should \f[B]never\f[] split long string literals like
+URLs or some other often copied entities over multiple lines to fit this
+limit:
+.PP
+Correct:
+.IP
+.nf
+\f[C]
+\[aq]https://www.youtube.com/watch?v=FqZTN594JQw&list=PLMYEtVRpaqY00V9W81Cwmzp6N6vZqfUKD4\[aq]
+\f[]
+.fi
+.PP
+Incorrect:
+.IP
+.nf
+\f[C]
+\[aq]https://www.youtube.com/watch?v=FqZTN594JQw&list=\[aq]
+\[aq]PLMYEtVRpaqY00V9W81Cwmzp6N6vZqfUKD4\[aq]
+\f[]
+.fi
+.SS Use safe conversion functions
+.PP
+Wrap all extracted numeric data into safe functions from
+\f[C]youtube_dl/utils.py\f[] (https://github.com/rg3/youtube-dl/blob/master/youtube_dl/utils.py):
+\f[C]int_or_none\f[], \f[C]float_or_none\f[].
+Use them for string to number conversions as well.
+.PP
+Use \f[C]url_or_none\f[] for safe URL processing.
+.PP
+Use \f[C]try_get\f[] for safe metadata extraction from parsed JSON.
+.PP
+Explore
+\f[C]youtube_dl/utils.py\f[] (https://github.com/rg3/youtube-dl/blob/master/youtube_dl/utils.py)
+for more useful convenience functions.
+.SS More examples
+.SS Safely extract optional description from parsed JSON
+.IP
+.nf
+\f[C]
+description\ =\ try_get(response,\ lambda\ x:\ x[\[aq]result\[aq]][\[aq]video\[aq]][0][\[aq]summary\[aq]],\ compat_str)
+\f[]
+.fi
+.SS Safely extract more optional metadata
+.IP
+.nf
+\f[C]
+video\ =\ try_get(response,\ lambda\ x:\ x[\[aq]result\[aq]][\[aq]video\[aq]][0],\ dict)\ or\ {}
+description\ =\ video.get(\[aq]summary\[aq])
+duration\ =\ float_or_none(video.get(\[aq]durationMs\[aq]),\ scale=1000)
+view_count\ =\ int_or_none(video.get(\[aq]views\[aq]))
+\f[]
+.fi