refactor: remove dead _get_channels_mapping code

[dangotbanned]

"_get_channels_mapping" is not accessed Pylance

I've been seeing this message for a while and thought I'd investigate.

This function appears in the first v5 commit to api

altair/altair/vegalite/v5/api.py

Lines 146 to 152 in 14bbb8d

	def _get_channels_mapping():
	mapping = {}
	for attr in dir(channels):
	cls = getattr(channels, attr)
	if isinstance(cls, type) and issubclass(cls, core.SchemaBase):
	mapping[cls] = attr.replace("Value", "").lower()
	return mapping

It was not referenced back then (2021), and still is not referenced.

It looks like an early version of the channels module manipulation code replaced in #3444

As this is a private function, I don't feel there's a strong case for keeping it around.
It also doesn't behave the same as the old infer_encoding_types - so any hypothetical usage wouldn't produce the same spec:

channel_objs = (getattr(channels, name) for name in dir(channels))
        if channels
    channel_objs = (
        else _ChannelCache.from_cache()
        c for c in channel_objs if isinstance(c, type) and issubclass(c, SchemaBase)
    )
    )
    channel_to_name: dict[type[SchemaBase], str] = {
        c: c._encoding_name for c in channel_objs
    }
    name_to_channel: dict[str, dict[str, type[SchemaBase]]] = {}
    for chan, name in channel_to_name.items():
        chans = name_to_channel.setdefault(name, {})
        if chan.__name__.endswith("Datum"):
            key = "datum"
        elif chan.__name__.endswith("Value"):
            key = "value"
        else:
            key = "field"
        chans[key] = chan

[joelostblom]

LGTM, but I'll leave it open in case someone else wants to confirm

[dangotbanned]

Think there was a false positive in https://github.com/vega/altair/actions/runs/9905108163/job/27363929771?pr=3467

Can't repro locally

>>> hatch test
cmd [1] | ruff check .
All checks passed!
cmd [2] | ruff format --diff --check .
84 files already formatted 
cmd [3] | mypy altair tests
Success: no issues found in 367 source files
cmd [4] | pytest -p no:randomly -n logical --numprocesses=logical --doctest-modules tests altair
==================================== 1257 passed, 1 skipped, 1 xfailed, 2 xpassed in 52.68s ====================================

[dangotbanned]

Think there was a false positive in vega/altair/actions/runs/9905108163/job/27363929771?pr=3467

Can't repro locally

>>> hatch test
cmd [1] | ruff check .
All checks passed!
cmd [2] | ruff format --diff --check .
84 files already formatted 
cmd [3] | mypy altair tests
Success: no issues found in 367 source files
cmd [4] | pytest -p no:randomly -n logical --numprocesses=logical --doctest-modules tests altair
==================================== 1257 passed, 1 skipped, 1 xfailed, 2 xpassed in 52.68s ====================================

I see this is failing again in https://github.com/vega/altair/actions/runs/9905108163/job/27364311474?pr=3467

@mattijn should I fix this in a new PR or in this one?

It isn't related to the changes I've made here at all

[mattijn]

Do we know where this error was introduced? I can't remember merging a PR with errors in CI.

[dangotbanned]

Do we know where this error was introduced? I can't remember merging a PR with errors in CI.

@mattijn

I've got a fix for it I'm about to push on a new PR #3475

It looks to be a change over at jsonschema, but I think it is unlikely to be python-jsonschema/jsonschema@0024b58#diff-0187567512f508c6badb2b034443e72b00316d8c0f40d5d7078eb5cb9adaf9f4