Scrapy-zyte-smartproxy

-------------------

Removed official support for Python 3.4, 3.5, 3.6, 3.7 and 3.8. Added official
Python 3.13 support.

`Backward-compatible
<https://docs.zyte.com/zyte-api/migration/zyte/smartproxy.html#spm-migrate-map>`__
``X-Crawlera``-prefixed headers are no longer translated into their matching
`Zyte API proxy mode headers
<https://docs.zyte.com/zyte-api/usage/proxy-mode.html#zapi-proxy-headers>`_,
Zyte API now handles their translation on the server side.

Added a new ``ZYTE_SMARTPROXY_KEEP_HEADERS`` setting that allows disabling
header dropping and translation.

-------------------

Ban and throttling responses from `Zyte API`_ `proxy mode`_ are now handled in
line with matching responses from Zyte Smart Proxy Manager.

-------------------

`Zyte API`_ `proxy mode`_ now has its own stat prefix.

Some user-facing messages mentioning only Zyte Smart Proxy Manager have also
been updated to reflect the fact that scrapy-zyte-smartproxy also supports Zyte
API proxy mode.

-------------------

Fix response handling for `Zyte API`_ `proxy mode`_. Before, a single
connection issue during a request would add a 90 second delay between requests
until the end of the crawl, instead of removing the delay after the first
successful response.

-------------------

Detect scenarios where the ``proxy`` ``Request.meta`` key has probably been
accidentally copied from an earlier response, warn about it, and fix the value.

The ``Zyte-Client`` header is again sent when using `Zyte API`_ `proxy mode`_,
now that Zyte API supports it.

-------------------

Fixed `Zyte API`_ `proxy mode`_ support by removing the mapping of unsupported
headers ``Zyte-Client`` and ``Zyte-No-Bancheck``.