pyodide-kernel Investigate further startup time improvements

Problem

pyodide_kernel requires many packages to be fully downloaded (if not already cached) and installed before giving the user the ability to do interactive computing.

Proposed Solution

Investigate additional packages which could be avoided, either by shim or patch.

Additional context

With jupyterlite/jupyterlite#913, we'd be down to 724 modules.

>>> import sys
>>> sorted(sys.modules)
['IPython',
 'IPython.core',
 'IPython.core.alias',
 'IPython.core.application',
 'IPython.core.async_helpers',
 'IPython.core.autocall',
 'IPython.core.builtin_trap',
 'IPython.core.compilerop',
 'IPython.core.completer',
 'IPython.core.completerlib',
 'IPython.core.crashhandler',
 'IPython.core.debugger',
 'IPython.core.display',
 'IPython.core.display_functions',
 'IPython.core.display_trap',
 'IPython.core.displayhook',
 'IPython.core.displaypub',
 'IPython.core.error',
 'IPython.core.events',
 'IPython.core.excolors',
 'IPython.core.extensions',
 'IPython.core.formatters',
 'IPython.core.getipython',
 'IPython.core.history',
 'IPython.core.hooks',
 'IPython.core.inputtransformer2',
 'IPython.core.interactiveshell',
 'IPython.core.latex_symbols',
 'IPython.core.logger',
 'IPython.core.macro',
 'IPython.core.magic',
 'IPython.core.magic_arguments',
 'IPython.core.magics',
 'IPython.core.magics.auto',
 'IPython.core.magics.basic',
 'IPython.core.magics.code',
 'IPython.core.magics.config',
 'IPython.core.magics.display',
 'IPython.core.magics.execution',
 'IPython.core.magics.extension',
 'IPython.core.magics.history',
 'IPython.core.magics.logging',
 'IPython.core.magics.namespace',
 'IPython.core.magics.osm',
 'IPython.core.magics.packaging',
 'IPython.core.magics.pylab',
 'IPython.core.magics.script',
 'IPython.core.oinspect',
 'IPython.core.page',
 'IPython.core.payload',
 'IPython.core.prefilter',
 'IPython.core.profiledir',
 'IPython.core.pylabtools',
 'IPython.core.release',
 'IPython.core.shellapp',
 'IPython.core.splitinput',
 'IPython.core.ultratb',
 'IPython.core.usage',
 'IPython.display',
 'IPython.extensions',
 'IPython.extensions.storemagic',
 'IPython.lib',
 'IPython.lib.clipboard',
 'IPython.lib.display',
 'IPython.lib.pretty',
 'IPython.paths',
 'IPython.terminal',
 'IPython.terminal.debugger',
 'IPython.terminal.embed',
 'IPython.terminal.interactiveshell',
 'IPython.terminal.ipapp',
 'IPython.terminal.magics',
 'IPython.terminal.prompts',
 'IPython.terminal.pt_inputhooks',
 'IPython.terminal.ptutils',
 'IPython.terminal.shortcuts',
 'IPython.testing',
 'IPython.testing.skipdoctest',
 'IPython.utils',
 'IPython.utils.PyColorize',
 'IPython.utils._process_common',
 'IPython.utils._process_posix',
 'IPython.utils._sysinfo',
 'IPython.utils.capture',
 'IPython.utils.colorable',
 'IPython.utils.coloransi',
 'IPython.utils.contexts',
 'IPython.utils.data',
 'IPython.utils.decorators',
 'IPython.utils.dir2',
 'IPython.utils.docs',
 'IPython.utils.encoding',
 'IPython.utils.frame',
 'IPython.utils.generics',
 'IPython.utils.importstring',
 'IPython.utils.io',
 'IPython.utils.ipstruct',
 'IPython.utils.module_paths',
 'IPython.utils.openpy',
 'IPython.utils.path',
 'IPython.utils.process',
 'IPython.utils.py3compat',
 'IPython.utils.sentinel',
 'IPython.utils.strdispatch',
 'IPython.utils.sysinfo',
 'IPython.utils.syspathcontext',
 'IPython.utils.terminal',
 'IPython.utils.text',
 'IPython.utils.timing',
 'IPython.utils.tokenutil',
 'IPython.utils.wildcard',
 '__future__',
 '__main__',
 '_abc',
 '_ast',
 '_bisect',
 '_blake2',
 '_bz2',
 '_codecs',
 '_collections',
 '_collections_abc',
 '_compat_pickle',
 '_compression',
 '_contextvars',
 '_csv',
 '_datetime',
 '_decimal',
 '_frozen_importlib',
 '_frozen_importlib_external',
 '_functools',
 '_heapq',
 '_imp',
 '_io',
 '_json',
 '_locale',
 '_lsprof',
 '_md5',
 '_operator',
 '_pickle',
 '_posixsubprocess',
 '_pyodide',
 '_pyodide._base',
 '_pyodide._core_docs',
 '_pyodide._importhook',
 '_pyodide.docstring',
 '_pyodide_core',
 '_queue',
 '_random',
 '_sha1',
 '_sha256',
 '_sha3',
 '_sha512',
 '_signal',
 '_sitebuiltins',
 '_socket',
 '_sqlite3',
 '_sre',
 '_stat',
 '_string',
 '_strptime',
 '_struct',
 '_sysconfigdata__emscripten_wasm32-emscripten',
 '_thread',
 '_warnings',
 '_weakref',
 '_weakrefset',
 'abc',
 'argparse',
 'array',
 'ast',
 'asttokens',
 'asttokens.asttokens',
 'asttokens.line_numbers',
 'asttokens.util',
 'asyncio',
 'asyncio.base_events',
 'asyncio.base_futures',
 'asyncio.base_subprocess',
 'asyncio.base_tasks',
 'asyncio.constants',
 'asyncio.coroutines',
 'asyncio.events',
 'asyncio.exceptions',
 'asyncio.format_helpers',
 'asyncio.futures',
 'asyncio.locks',
 'asyncio.log',
 'asyncio.mixins',
 'asyncio.protocols',
 'asyncio.queues',
 'asyncio.runners',
 'asyncio.selector_events',
 'asyncio.sslproto',
 'asyncio.staggered',
 'asyncio.streams',
 'asyncio.subprocess',
 'asyncio.tasks',
 'asyncio.threads',
 'asyncio.transports',
 'asyncio.trsock',
 'asyncio.unix_events',
 'atexit',
 'backcall',
 'backcall.backcall',
 'base64',
 'bdb',
 'binascii',
 'bisect',
 'builtins',
 'bz2',
 'cProfile',
 'calendar',
 'cmd',
 'code',
 'codecs',
 'codeop',
 'collections',
 'collections.abc',
 'colorsys',
 'concurrent',
 'concurrent.futures',
 'concurrent.futures._base',
 'concurrent.futures.thread',
 'contextlib',
 'contextvars',
 'copy',
 'copyreg',
 'csv',
 'dataclasses',
 'datetime',
 'decimal',
 'decorator',
 'difflib',
 'dis',
 'email',
 'email._encoded_words',
 'email._parseaddr',
 'email._policybase',
 'email.base64mime',
 'email.charset',
 'email.encoders',
 'email.errors',
 'email.feedparser',
 'email.header',
 'email.iterators',
 'email.message',
 'email.parser',
 'email.quoprimime',
 'email.utils',
 'encodings',
 'encodings.aliases',
 'encodings.cp437',
 'encodings.utf_8',
 'enum',
 'errno',
 'executing',
 'executing.executing',
 'executing.version',
 'fcntl',
 'filecmp',
 'fnmatch',
 'fractions',
 'functools',
 'gc',
 'genericpath',
 'getopt',
 'getpass',
 'gettext',
 'glob',
 'gzip',
 'hashlib',
 'heapq',
 'html',
 'html.entities',
 'http',
 'http.client',
 'importlib',
 'importlib._abc',
 'importlib._bootstrap',
 'importlib._bootstrap_external',
 'importlib.abc',
 'importlib.machinery',
 'importlib.metadata',
 'importlib.metadata._adapters',
 'importlib.metadata._collections',
 'importlib.metadata._functools',
 'importlib.metadata._itertools',
 'importlib.metadata._meta',
 'importlib.metadata._text',
 'importlib.util',
 'inspect',
 'io',
 'ipykernel',
 'ipykernel.comm',
 'ipykernel.jsonutil',
 'itertools',
 'jedi',
 'jedi._compatibility',
 'jedi.api',
 'jedi.api.classes',
 'jedi.api.completion',
 'jedi.api.completion_cache',
 'jedi.api.environment',
 'jedi.api.errors',
 'jedi.api.exceptions',
 'jedi.api.file_name',
 'jedi.api.helpers',
 'jedi.api.interpreter',
 'jedi.api.keywords',
 'jedi.api.project',
 'jedi.api.refactoring',
 'jedi.api.refactoring.extract',
 'jedi.api.strings',
 'jedi.cache',
 'jedi.common',
 'jedi.debug',
 'jedi.file_io',
 'jedi.inference',
 'jedi.inference.analysis',
 'jedi.inference.arguments',
 'jedi.inference.base_value',
 'jedi.inference.cache',
 'jedi.inference.compiled',
 'jedi.inference.compiled.access',
 'jedi.inference.compiled.getattr_static',
 'jedi.inference.compiled.mixed',
 'jedi.inference.compiled.subprocess',
 'jedi.inference.compiled.subprocess.functions',
 'jedi.inference.compiled.value',
 'jedi.inference.context',
 'jedi.inference.docstring_utils',
 'jedi.inference.docstrings',
 'jedi.inference.filters',
 'jedi.inference.flow_analysis',
 'jedi.inference.gradual',
 'jedi.inference.gradual.annotation',
 'jedi.inference.gradual.base',
 'jedi.inference.gradual.conversion',
 'jedi.inference.gradual.generics',
 'jedi.inference.gradual.stub_value',
 'jedi.inference.gradual.type_var',
 'jedi.inference.gradual.typeshed',
 'jedi.inference.gradual.typing',
 'jedi.inference.gradual.utils',
 'jedi.inference.helpers',
 'jedi.inference.imports',
 'jedi.inference.lazy_value',
 'jedi.inference.names',
 'jedi.inference.param',
 'jedi.inference.parser_cache',
 'jedi.inference.recursion',
 'jedi.inference.references',
 'jedi.inference.signature',
 'jedi.inference.syntax_tree',
 'jedi.inference.sys_path',
 'jedi.inference.utils',
 'jedi.inference.value',
 'jedi.inference.value.decorator',
 'jedi.inference.value.dynamic_arrays',
 'jedi.inference.value.function',
 'jedi.inference.value.instance',
 'jedi.inference.value.iterable',
 'jedi.inference.value.klass',
 'jedi.inference.value.module',
 'jedi.parser_utils',
 'jedi.plugins',
 'jedi.plugins.django',
 'jedi.plugins.flask',
 'jedi.plugins.pytest',
 'jedi.plugins.registry',
 'jedi.plugins.stdlib',
 'jedi.settings',
 'js',
 'json',
 'json.decoder',
 'json.encoder',
 'json.scanner',
 'keyword',
 'linecache',
 'locale',
 'logging',
 'logging.config',
 'logging.handlers',
 'marshal',
 'math',
 'micropip',
 'micropip._compat',
 'micropip._compat_in_pyodide',
 'micropip._micropip',
 'micropip.externals',
 'micropip.externals.pip',
 'micropip.externals.pip._internal',
 'micropip.externals.pip._internal.utils',
 'micropip.externals.pip._internal.utils.pkg_resources',
 'micropip.externals.pip._internal.utils.wheel',
 'micropip.externals.pip._vendor',
 'micropip.externals.pip._vendor.pkg_resources',
 'micropip.package',
 'mimetypes',
 'ntpath',
 'numbers',
 'opcode',
 'operator',
 'os',
 'os.path',
 'packaging',
 'packaging.__about__',
 'packaging._manylinux',
 'packaging._musllinux',
 'packaging._structures',
 'packaging.markers',
 'packaging.requirements',
 'packaging.specifiers',
 'packaging.tags',
 'packaging.utils',
 'packaging.version',
 'parso',
 'parso._compatibility',
 'parso.cache',
 'parso.file_io',
 'parso.grammar',
 'parso.normalizer',
 'parso.parser',
 'parso.pgen2',
 'parso.pgen2.generator',
 'parso.pgen2.grammar_parser',
 'parso.python',
 'parso.python.diff',
 'parso.python.errors',
 'parso.python.parser',
 'parso.python.pep8',
 'parso.python.prefix',
 'parso.python.token',
 'parso.python.tokenize',
 'parso.python.tree',
 'parso.tree',
 'parso.utils',
 'pathlib',
 'pdb',
 'pexpect',
 'pickle',
 'pickleshare',
 'piplite',
 'piplite.piplite',
 'pkgutil',
 'platform',
 'posix',
 'posixpath',
 'pprint',
 'profile',
 'prompt_toolkit',
 'prompt_toolkit.application',
 'prompt_toolkit.application.application',
 'prompt_toolkit.application.current',
 'prompt_toolkit.application.dummy',
 'prompt_toolkit.application.run_in_terminal',
 'prompt_toolkit.auto_suggest',
 'prompt_toolkit.buffer',
 'prompt_toolkit.cache',
 'prompt_toolkit.clipboard',
 'prompt_toolkit.clipboard.base',
 'prompt_toolkit.clipboard.in_memory',
 'prompt_toolkit.completion',
 'prompt_toolkit.completion.base',
 'prompt_toolkit.completion.deduplicate',
 'prompt_toolkit.completion.filesystem',
 'prompt_toolkit.completion.fuzzy_completer',
 'prompt_toolkit.completion.nested',
 'prompt_toolkit.completion.word_completer',
 'prompt_toolkit.cursor_shapes',
 'prompt_toolkit.data_structures',
 'prompt_toolkit.document',
 'prompt_toolkit.enums',
 'prompt_toolkit.eventloop',
 'prompt_toolkit.eventloop.async_context_manager',
 'prompt_toolkit.eventloop.async_generator',
 'prompt_toolkit.eventloop.inputhook',
 'prompt_toolkit.eventloop.utils',
 'prompt_toolkit.filters',
 'prompt_toolkit.filters.app',
 'prompt_toolkit.filters.base',
 'prompt_toolkit.filters.cli',
 'prompt_toolkit.filters.utils',
 'prompt_toolkit.formatted_text',
 'prompt_toolkit.formatted_text.ansi',
 'prompt_toolkit.formatted_text.base',
 'prompt_toolkit.formatted_text.html',
 'prompt_toolkit.formatted_text.pygments',
 'prompt_toolkit.formatted_text.utils',
 'prompt_toolkit.history',
 'prompt_toolkit.input',
 'prompt_toolkit.input.ansi_escape_sequences',
 'prompt_toolkit.input.base',
 'prompt_toolkit.input.defaults',
 'prompt_toolkit.input.typeahead',
 'prompt_toolkit.input.vt100_parser',
 'prompt_toolkit.key_binding',
 'prompt_toolkit.key_binding.bindings',
 'prompt_toolkit.key_binding.bindings.auto_suggest',
 'prompt_toolkit.key_binding.bindings.basic',
 'prompt_toolkit.key_binding.bindings.completion',
 'prompt_toolkit.key_binding.bindings.cpr',
 'prompt_toolkit.key_binding.bindings.emacs',
 'prompt_toolkit.key_binding.bindings.focus',
 'prompt_toolkit.key_binding.bindings.mouse',
 'prompt_toolkit.key_binding.bindings.named_commands',
 'prompt_toolkit.key_binding.bindings.open_in_editor',
 'prompt_toolkit.key_binding.bindings.page_navigation',
 'prompt_toolkit.key_binding.bindings.scroll',
 'prompt_toolkit.key_binding.bindings.vi',
 'prompt_toolkit.key_binding.defaults',
 'prompt_toolkit.key_binding.digraphs',
 'prompt_toolkit.key_binding.emacs_state',
 'prompt_toolkit.key_binding.key_bindings',
 'prompt_toolkit.key_binding.key_processor',
 'prompt_toolkit.key_binding.vi_state',
 'prompt_toolkit.keys',
 'prompt_toolkit.layout',
 'prompt_toolkit.layout.containers',
 'prompt_toolkit.layout.controls',
 'prompt_toolkit.layout.dimension',
 'prompt_toolkit.layout.dummy',
 'prompt_toolkit.layout.layout',
 'prompt_toolkit.layout.margins',
 'prompt_toolkit.layout.menus',
 'prompt_toolkit.layout.mouse_handlers',
 'prompt_toolkit.layout.processors',
 'prompt_toolkit.layout.screen',
 'prompt_toolkit.layout.scrollable_pane',
 'prompt_toolkit.layout.utils',
 'prompt_toolkit.lexers',
 'prompt_toolkit.lexers.base',
 'prompt_toolkit.lexers.pygments',
 'prompt_toolkit.mouse_events',
 'prompt_toolkit.output',
 'prompt_toolkit.output.base',
 'prompt_toolkit.output.color_depth',
 'prompt_toolkit.output.defaults',
 'prompt_toolkit.output.flush_stdout',
 'prompt_toolkit.output.plain_text',
 'prompt_toolkit.output.vt100',
 'prompt_toolkit.patch_stdout',
 'prompt_toolkit.renderer',
 'prompt_toolkit.search',
 'prompt_toolkit.selection',
 'prompt_toolkit.shortcuts',
 'prompt_toolkit.shortcuts.dialogs',
 'prompt_toolkit.shortcuts.progress_bar',
 'prompt_toolkit.shortcuts.progress_bar.base',
 'prompt_toolkit.shortcuts.progress_bar.formatters',
 'prompt_toolkit.shortcuts.prompt',
 'prompt_toolkit.shortcuts.utils',
 'prompt_toolkit.styles',
 'prompt_toolkit.styles.base',
 'prompt_toolkit.styles.defaults',
 'prompt_toolkit.styles.named_colors',
 'prompt_toolkit.styles.pygments',
 'prompt_toolkit.styles.style',
 'prompt_toolkit.styles.style_transformation',
 'prompt_toolkit.utils',
 'prompt_toolkit.validation',
 'prompt_toolkit.widgets',
 'prompt_toolkit.widgets.base',
 'prompt_toolkit.widgets.dialogs',
 'prompt_toolkit.widgets.menus',
 'prompt_toolkit.widgets.toolbars',
 'pstats',
 'pure_eval',
 'pure_eval.core',
 'pure_eval.my_getattr_static',
 'pure_eval.utils',
 'pure_eval.version',
 'pydoc',
 'pydoc_data',
 'pydoc_data.topics',
 'pygments',
 'pygments.console',
 'pygments.filter',
 'pygments.filters',
 'pygments.formatter',
 'pygments.formatters',
 'pygments.formatters._mapping',
 'pygments.formatters.html',
 'pygments.formatters.terminal256',
 'pygments.lexer',
 'pygments.lexers',
 'pygments.lexers._mapping',
 'pygments.lexers.python',
 'pygments.modeline',
 'pygments.plugin',
 'pygments.regexopt',
 'pygments.style',
 'pygments.styles',
 'pygments.token',
 'pygments.unistring',
 'pygments.util',
 'pyodide',
 'pyodide._core',
 'pyodide._package_loader',
 'pyodide._state',
 'pyodide.code',
 'pyodide.ffi',
 'pyodide.http',
 'pyodide.webloop',
 'pyodide_js',
 'pyodide_js._api',
 'pyolite',
 'pyolite.display',
 'pyolite.interpreter',
 'pyolite.kernel',
 'pyolite.litetransform',
 'pyolite.mocks',
 'pyolite.patches',
 'pyparsing',
 'pyparsing.actions',
 'pyparsing.common',
 'pyparsing.core',
 'pyparsing.exceptions',
 'pyparsing.helpers',
 'pyparsing.results',
 'pyparsing.testing',
 'pyparsing.unicode',
 'pyparsing.util',
 'queue',
 'quopri',
 'random',
 're',
 'reprlib',
 'resource',
 'runpy',
 'select',
 'selectors',
 'shlex',
 'shutil',
 'signal',
 'site',
 'six',
 'six.moves',
 'socket',
 'socketserver',
 'sqlite3',
 'sqlite3.dbapi2',
 'sre_compile',
 'sre_constants',
 'sre_parse',
 'stack_data',
 'stack_data.core',
 'stack_data.formatting',
 'stack_data.serializing',
 'stack_data.utils',
 'stack_data.version',
 'stat',
 'string',
 'struct',
 'subprocess',
 'sys',
 'sysconfig',
 'tarfile',
 'tempfile',
 'termios',
 'textwrap',
 'threading',
 'time',
 'timeit',
 'token',
 'tokenize',
 'tornado',
 'tornado.gen',
 'traceback',
 'traitlets',
 'traitlets._version',
 'traitlets.config',
 'traitlets.config.application',
 'traitlets.config.configurable',
 'traitlets.config.loader',
 'traitlets.traitlets',
 'traitlets.utils',
 'traitlets.utils.bunch',
 'traitlets.utils.decorators',
 'traitlets.utils.descriptions',
 'traitlets.utils.getargspec',
 'traitlets.utils.importstring',
 'traitlets.utils.nested_update',
 'traitlets.utils.sentinel',
 'traitlets.utils.text',
 'types',
 'typing',
 'typing.io',
 'typing.re',
 'unicodedata',
 'unittest',
 'unittest.case',
 'unittest.loader',
 'unittest.main',
 'unittest.mock',
 'unittest.result',
 'unittest.runner',
 'unittest.signals',
 'unittest.suite',
 'unittest.util',
 'urllib',
 'urllib.error',
 'urllib.parse',
 'urllib.request',
 'urllib.response',
 'uu',
 'uuid',
 'warnings',
 'wcwidth',
 'wcwidth.table_wide',
 'wcwidth.table_zero',
 'wcwidth.unicode_versions',
 'wcwidth.wcwidth',
 'weakref',
 'xml',
 'xml.dom',
 'xml.dom.NodeFilter',
 'xml.dom.domreg',
 'xml.dom.minicompat',
 'xml.dom.minidom',
 'xml.dom.xmlbuilder',
 'zipfile',
 'zipimport',
 'zlib']

A standout are the 97 jedi and parso packages: at present, jedi integration is turned off, presumably as it is so slow as to appear broken. At any rate, we're still paying 1.7mb on the wire to download these two packages.

As jedi is part of the pyodide standard distribution it may be rather hacky to work around it getting installed, due to issues like jupyterlite/jupyterlite#904, but may be worth it for doign 10% less work for... nothing.

Dec 13 '22 15:12 bollwyvl

As jedi is part of the pyodide standard distribution it may be rather hacky to work around it getting installed, due to issues like jupyterlite/jupyterlite#904, but may be worth it for doign 10% less work for... nothing.

I saw a comment like this in jupyterlite/jupyterlite#911 as well, but unless I'm misunderstanding how pyolite and pyodide interact, I don't think a package being part of the 'pyodide standard distribution' means that it gets installed automatically. The only reason things like import matplotlib or import jedi work in the pyodide online console is that, similarly to jupyterlite, it parses the imports and automatically installs standard packages that are being imported:

https://github.com/pyodide/pyodide/blob/97cd5bdc1cf62f4e5e44a305a7682d92b556a1e0/src/py/pyodide/console.py#L478

Dec 13 '22 16:12 jobovy

it parses the imports and automatically installs standard packages that are being imported:

Right, iirc even code like the following would trigger the install?

if False:
    import matplotlib

Dec 13 '22 16:12 jtpio

Right: unless it's made optional upstream in IPython (unlikely), the approach we'd need today:

create a stub package that satisfied the import jedi in IPython
- or not, as it's already conditional and ready to fail
patch pyodide's knowledge of its packages really early
- as this is inside the worker, our fetch hack wouldn't work

This would theoretically be something we could hoist up to lite plugin settings, e.g.

"jupyter-config-data": {
  "litePluginSettings": {
    "@jupyterlite/pyolite-kernel-extension:kernel": {
      "disablePyPIFallback": true,
      "repodata": {
         "jedi": null
      }
    }
  }
}

this would then allow piplite to answer the response for a pre-packaged wheel.

Dec 13 '22 16:12 bollwyvl

Right, iirc even code like the following would trigger the install?
if False:
    import matplotlib

Yes, I think it purely, statically looks forimport statements and imports everything it finds that's in the standard distribution:

https://github.com/pyodide/pyodide/blob/761b6320ba6a795630c4d44df9f253fd0cd0c562/src/py/_pyodide/_base.py#L514

Dec 13 '22 16:12 jobovy

patch pyodide's knowledge of its packages really early

as this is inside the worker, our fetch hack wouldn't work

I think the only reason you install pyodide's version is because you run code through loadPackagesFromImports before running it, no? So you could just remove any package you don't want to install from pyodide from the list to import, like I did here to fix jupyterlite/jupyterlite#904:

https://github.com/jobovy/jupyterlite/compare/quiet-prerun...quiet-prerun-purge-galpy

Dec 13 '22 16:12 jobovy

Right, but 99% of the time, the user wants that feature.

Dec 13 '22 16:12 bollwyvl

Right, but 99% of the time, the user wants that feature.

Yes, that's why I just remove galpy from the list in my case and still automatically install everything else. Here you could similarly remove jedi from the list of packages to ever install from pyodide and install your shim instead?

Dec 13 '22 16:12 jobovy

Right, except this is for the kernel itself, so there are some timing issues. We further allow folk to use a custom pyodide, and don't/won't ship all 200mb.

Typically if we do some hack like this, we try to make it useful for downstreams as well (this is why we have piplite in the first place, to decouple the names of packages from the typescript code... almost)

Dec 13 '22 17:12 bollwyvl

While I was thinking this over:

pyodide's repodata.json has a critical advantage vs its conda namesake and the Warehouse JSON API, namely that it includes the imports supported by the distribution (e.g. pillow provides PIL)
we already do "wheel stuff" with as-downloaded wheels with pkginfo to generate one JSON file
we could just as well generate a pyodide-compatible repodata.json, and (basically) extract the names inside the packages
this would mean that pre-indexed wheels would not need to be pip installed, they would get picked up by the auto-import mechanism, without sacrificing debuggability when things fail

This wouldn't change the public API at all, and notebooks would still have to import whatever, but is probably enough of a carrot to look into how we might solve both the jedi patch, while encouraging site owners to provide well-known versions of the wheels their content will use.

Dec 14 '22 21:12 bollwyvl