Skip to content

feat(local-dev): pick a usable Python and offer to install a missing toolchain - #7073

Closed
Yicong-Huang wants to merge 1 commit into
fix/local-dev-linuxfrom
feat/local-dev-toolchain-install
Closed

feat(local-dev): pick a usable Python and offer to install a missing toolchain#7073
Yicong-Huang wants to merge 1 commit into
fix/local-dev-linuxfrom
feat/local-dev-toolchain-install

Conversation

@Yicong-Huang

Copy link
Copy Markdown
Contributor

Stacked on #7070 (fix/local-dev-linux), which it needs for the
platform detection. Review that one first; this PR's own diff is the second
commit. I'll rebase onto main once #7070 lands.

What changes were proposed in this PR?

Two halves of the same fresh-machine problem.

The interpreter that runs Python UDFs was picked blindly.
UDF_PYTHON_PATH defaulted to command -v python3 at source time — the system
interpreter, which almost never has amber/requirements.txt installed. Python
UDFs then failed at worker launch on a stack of import errors that pointed
nowhere near the interpreter choice. The venv AGENTS.md tells contributors to
create (<workspace>/venv312) was ignored, as was an already-activated
$VIRTUAL_ENV.

Resolution is now lazy, and takes the first candidate that is Python 3.12
and can import what a worker needs:

$UDF_PYTHON_PATH  ->  $VIRTUAL_ENV  ->  <workspace>/venv312  ->  pyenv 3.12
  ->  python3.12 on PATH  ->  python3 on PATH

The choice is reported the way the JDK probe reports JAVA_HOME. An explicit
UDF_PYTHON_PATH still wins — it is the documented override — but now says so
when it can't import the deps. "No 3.12 anywhere" and "3.12 without amber's
deps" are reported differently, because they need different fixes and the old
default silently produced the second one. Lazy so status / logs / --help
don't spawn an interpreter per candidate, and never fatal: the JVM services run
fine without a Python toolchain, only UDFs don't.

A missing tool was described, never offered. _install_hint printed a
suggestion and gave up, so a fresh machine meant hand-installing five or six
things, re-running up after each to discover the next gap. A missing JDK 17,
Node or Python 3.12 is now offered for install, after showing the exact command:

Tool Installer
JDK 17 distro package manager (apt-get/dnf/yum/pacman/zypper), brew on macOS, SDKMAN when there is none
Node 24 nvm (bootstraps nvm first if absent)
Python 3.12 pyenv (bootstraps pyenv first if absent)

jenv is not used: it switches between JDKs, it cannot install one.

The decision helpers (_pkg_manager, _install_cmd_for) only ever print
what would run; a separate step executes it. That keeps the choice
unit-testable without installing anything, and it means the user sees the
command — sudo included — before being asked.

Behaviour deliberately preserved for non-interactive callers: without a TTY
nothing is ever prompted
and the old print-a-hint-and-fail path stands, so
scripted and CI runs cannot hang on a read. --install-missing answers yes
without asking, --no-install only ever prints the hint, and the contradiction
is refused rather than silently resolved. docker and sbt stay hint-only —
docker needs a daemon, group membership and a re-login, which is not something
to do behind a y/n prompt.

Any related issues, documentation, discussions?

Closes #7066
Stacked on #7070 (#7065).

How was this PR tested?

New coverage in the existing infra-job suite. Everything goes through the
pure decision helpers, so CI asserts on what would be installed without
installing anything; the Python probes are driven by fake interpreters that
echo a version or set an exit code:

$ bash bin/local-dev/tests/test_local_dev_sh.sh
...
  ✓ pkg manager: apt-get → apt-get
  ✓ pkg manager: dnf only → dnf
  ✓ pkg manager: pacman only → pacman
  ✓ pkg manager: apt-get preferred over dnf → apt-get
  ✓ pkg manager: nothing installed (refuses)
  ✓ _pkg_manager keeps the Darwin/brew branch
  ✓ install cmd: java via apt names openjdk-17-jdk
  ✓ install cmd: java via apt asks for sudo explicitly
  ✓ install cmd: java via dnf names java-17-openjdk-devel
  ✓ install cmd: java with no package manager falls back to SDKMAN
  ✓ install cmd: node uses nvm
  ✓ install cmd: python uses pyenv
  ✓ install cmd: python pins 3.12
  ✓ install cmd: refuses 'docker'
  ✓ install cmd: refuses 'sbt'
  ✓ install cmd: refuses 'definitely-not-a-tool'
  ✓ consent: refuses without a TTY (never prompts)
  ✓ consent: TEXERA_INSTALL_MISSING=1 assumes yes
  ✓ consent: TEXERA_INSTALL_MISSING=0 refuses
  ✓ python probe: 3.12 accepted
  ✓ python probe: 3.11 rejected
  ✓ python probe: missing interpreter rejected
  ✓ python probe: importable deps accepted
  ✓ python probe: missing deps rejected
  ✓ python probe: empty path rejected
  ✓ python candidates: $VIRTUAL_ENV before sibling venv312
  ✓ python candidates: venv312 before bare python3 on PATH
  ✓ _require_udf_python helper is defined
  ✓ cmd_up resolves the UDF interpreter before launching
  ✓ cmd_auto resolves the UDF interpreter before launching
  ✓ cmd_up_one resolves the UDF interpreter before launching
  ✓ _require_udf_python points at amber's requirements, not the TUI hint
  ✓ UDF_PYTHON_PATH no longer defaults to bare python3 at source time
  ✓ --help documents --install-missing / --no-install
  ✓ up rejects --install-missing together with --no-install (rc=2)

103 passed, 0 failed
$ python -m pytest bin/local-dev/tests/ -q
1 failed, 42 passed

test_is_dirty_after_seed_then_edit is pre-existing and unrelated — it
reproduces identically on a pristine upstream/main checkout on this machine,
and this PR touches neither tui.py nor that test. Filed separately.

Manually, on Ubuntu 24.04.4 with the amber deps installed into a sibling
venv312 and nothing in UDF_PYTHON_PATH — it finds the venv rather than the
system interpreter the old default would have taken:

$ bin/local-dev.sh auto
  ✓  python: /home/…/Repos/venv312/bin/python  (runs Python UDFs)

$ command -v python3            # what the old default resolved to
/usr/bin/python3
$ /usr/bin/python3 -c "import pyarrow"
ModuleNotFoundError: No module named 'pyarrow'

The consent path, forced by pinning a version that cannot exist
(TEXERA_PYTHON_VERSION=3.99). With a TTY it shows the command and asks; the
answer here was n, and the run continued:

  ⚠  python: no Python 3.99 found — Python UDFs will not run
  looked in: $VIRTUAL_ENV, <workspace>/venv312, pyenv, PATH
  or point at one yourself: export UDF_PYTHON_PATH=/path/to/python3.99

  python is missing. I can run:

      curl -fsSL https://pyenv.run | bash && export PYENV_ROOT="$HOME/.pyenv" && … && pyenv install -s 3.99

  Run it now? [y/N]

Without a TTY (</dev/null) the same run prints the warning, never prompts,
does not hang, and exits 0. Both flags are accepted on their own
(auto --no-install, auto --install-missing → rc 0) and rejected together
(rc 2).

Not exercised end-to-end: actually running the JDK / Node / pyenv installers,
since this machine already has all three. The commands themselves are asserted
by the tests above; the execution step is a bash -lc of exactly the string
shown to the user.

Was this PR authored or co-authored using generative AI tooling?

Generated-by: Claude Code (Opus 5)

@github-actions

Copy link
Copy Markdown
Contributor

Automated Reviewer Suggestions

Based on the git blame history of the changed files, we recommend the following reviewers:

  • No candidates found from git blame history.

@Yicong-Huang
Yicong-Huang force-pushed the fix/local-dev-linux branch from 4e2fcdf to 347c887 Compare July 29, 2026 22:44
@Yicong-Huang
Yicong-Huang force-pushed the feat/local-dev-toolchain-install branch from 99a3e30 to 78e091f Compare July 29, 2026 22:45
…toolchain

Two halves of the same fresh-machine problem.

**The interpreter that runs Python UDFs was picked blindly.**
`UDF_PYTHON_PATH` defaulted to `command -v python3` at source time — the system
interpreter, which almost never has `amber/requirements.txt` installed. Python
UDFs then failed at worker launch on a stack of import errors that pointed
nowhere near the interpreter choice, and the venv AGENTS.md tells contributors
to create (`<workspace>/venv312`) was ignored, as was an already-activated
`$VIRTUAL_ENV`.

Now resolved lazily, taking the first candidate that is Python 3.12 *and* can
import the worker's deps:

    $UDF_PYTHON_PATH -> $VIRTUAL_ENV -> <workspace>/venv312 -> pyenv 3.12
      -> python3.12 on PATH -> python3 on PATH

The chosen interpreter is reported the way the JDK probe reports JAVA_HOME. An
explicit `UDF_PYTHON_PATH` still wins — it's the documented override — but now
says so when it can't import what a worker needs. "No 3.12 anywhere" and "3.12
without amber's deps" are reported differently, because they need different
fixes and the old default silently produced the second one. Resolution is lazy
so `status` / `logs` / `--help` don't spawn an interpreter per candidate, and
never fatal: the JVM services run fine without a Python toolchain.

**A missing tool was described, never offered.** `_install_hint` printed a
suggestion and gave up, so a fresh machine meant hand-installing several things
and re-running `up` after each one to find the next gap. A missing JDK 17, Node
or Python 3.12 is now offered for install after showing the exact command:
the distro package manager for the JDK (SDKMAN when there is none), nvm for
Node, pyenv for Python — both bootstrapping themselves if absent.

The decision helpers (`_pkg_manager`, `_install_cmd_for`) only ever print what
would run; a separate step executes it. That keeps the choice unit-testable
without installing anything, and means the user is shown the command — `sudo`
included — before being asked.

Deliberately unchanged for non-interactive callers: without a TTY nothing is
ever prompted and the old print-a-hint-and-fail behaviour stands, so scripted
and CI runs can't hang on a read. `--install-missing` answers yes without
asking, `--no-install` only ever prints; the contradiction is refused rather
than silently resolved. docker and sbt stay hint-only — docker needs a daemon,
group membership and a re-login, which isn't something to do behind a y/n
prompt.

Closes #7066
@Yicong-Huang
Yicong-Huang force-pushed the fix/local-dev-linux branch from 347c887 to b68bfe3 Compare July 29, 2026 22:53
@Yicong-Huang
Yicong-Huang force-pushed the feat/local-dev-toolchain-install branch from 78e091f to 6025b63 Compare July 29, 2026 22:53
@Yicong-Huang

Copy link
Copy Markdown
Contributor Author

Superseded by #7078 — same commits, but opened from a fork branch as it should have been.

@Yicong-Huang
Yicong-Huang deleted the feat/local-dev-toolchain-install branch July 29, 2026 23:03
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

local-dev should detect the toolchain and offer to install what is missing

1 participant