[python] expose new encoder parameters as kwargs of brotli.compress

remove additional low-level parameters

modified the help string to include the new parameters.

renamed variables: transform -> enable_transforms, dictionary -> enable_dictionary

update brotli.compress docstring accordingly

Add a MODE_GENERIC compression mode to the interface.

With this the users can distinguish between not knowing
what the input is (ddefault) and knowing that it is text,
and thus can be relied on to force some UTF-8 specific settings.

[python] fix compatibility_test.py with Python 3

In python3, the 'string' module no longer has a 'split' function.

Handle multiple compressed files per original in the test.

Add some more test cases that decompress to the empty
file or a one byte long file. These test cases have
examples for the updated stream header and meta-block
header formats.

Restrict the ARM optimizations to little endian architectures.

Add an Acknowledgements section to the spec.

Fix 32bit build.

Support window bits 10 - 15 in the decoder.

The previous window bit value 17 is used to
extend the range, since it has not been used
in any previous encoders.

Align distance code meaning in the brotli encoder.

Two different definitions (offset by 1) were used in
command.h and hash.h. Now they have been made the same,
also consistent with the spec (e.g. 0 means use previous dist, etc...)

Faster encoding for low quality settings.

With this commit, the encoder will skip some
compression optimization steps for quality <= 4,
which results in faster compression but higher
compressed sizes.

Use a static context map with two buckets for UTF8 data.

Enabled for quality >= 4, and if there are no obvious
UTF8 violations detected.
For each block, we gather two separate histograms, one
for continuation bytes and one for ASCII or lead bytes.

Decoder optimizations for ARM architecture.

Speed and memory usage improvements for the decoder.

* Change order of members of bit reader state structure.

* Remove unused includes for assert. Add BROTLI_DCHECK
  macros and use it instead of assert.

* Do not calculate nbits in common case of ReadSymbol.

* Introduce and use PREDICT_TRUE / PREDICT_FALSE macros.

* Allocate less memory in the brotli decoder if it knows
  the result size beforehand. Before this, the decoder
  would always allocate 16MB if the encoder annotated the
  window size as 22 bit (which is the default), even if the
  file is only a few KB uncompressed. Now, it'll only
  allocate a ringbuffer as large as needed for the result file.
  But only if it can know the filesize, it's not possible
  to know that if there are multiple metablocks or too large
  uncompressed metablock.

Don't do any block splitting for quality 1.

Fix an error propagation bug in the decoder.