Files
LEDMatrix/src/plugin_system/plugin_loader.py
T
0c5b9c57d3 fix: keep low-memory boards reachable under load (#464)
* fix(service): survive corrupt health cache and clean exits

Three independent failure modes that each end with a dark panel and no
automatic recovery.

1. PluginHealthTracker._load_health_state returned the cached value
   verbatim. If that value is not a dict, every caller raises
   AttributeError: 'list' object has no attribute 'get' — during
   DisplayController.__init__, so the process dies before the display
   loop starts. systemd restarts it, the same bad entry is read back
   from disk, and it dies again: an unattended restart loop that
   survives reboots because the cause is persisted. Observed in the
   field with plugin_health:<id> holding an unrelated plugin's list
   payload. Now non-dict entries are discarded with a warning and the
   defaults are rebuilt.

2. ledmatrix.service used Restart=on-failure, so any exit with status 0
   left the unit stopped and the panel dark indefinitely — systemd
   treats it as success and never brings it back. Restart=always.

3. ledmatrix-wifi-monitor.service used StandardOutput=syslog, which
   systemd has marked obsolete; it warns and rewrites it to journal on
   every load.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* perf(memory): size the cache to the board and stop reinstalling deps

On a 1GB Pi 3B+ the display process settles around 600MB RSS of 905MB
total. When the remaining headroom runs out the failure is not a clean
crash: fork() starts returning ENOMEM, so sshd accepts connections and
closes them before its banner, timer jobs stop running, and the panel
goes dark, while already-resident processes keep serving normally. The
board looks healthy from outside and cannot be logged into. Only a power
cycle clears it.

Three contributing causes:

- MemoryCache had a fixed 1000-entry ceiling. Entries are parsed API
  payloads of tens of KB, so one ceiling cannot serve both a 512MB Zero
  2 W and an 8GB Pi 5. Now scaled from MemTotal (150 entries at <=1GB,
  1500 at >=8GB), overridable with LEDMATRIX_CACHE_MAX_ENTRIES.

- requirements_are_satisfied() returned False for any requirement with
  extras, so a plugin depending on python-socketio[client] re-ran pip on
  every single start: ~8s, a network dependency, and a 100-200MB spike
  at the least convenient moment. During a restart loop it repeats for
  each restart. Extras are now resolved one level deep against installed
  metadata, keeping the conservative "anything unverifiable falls
  through to pip" contract.

- ledmatrix.service had no memory ceiling. MemoryMax=85% expressed as a
  percentage so one unit file suits every board. Note this needs the
  memory cgroup controller, which Pi firmware disables by default;
  first_time_install.sh now adds cgroup_enable=memory to cmdline.txt,
  and the unit file documents how to verify it took effect.

first_time_install.sh also enables persistent journald storage (capped
at 64M). Default storage is volatile, so every reboot destroys the logs
that would explain why the board rebooted.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* docs: guidance for 512MB and 1GB boards

Documents the memory ceiling on small boards and, more usefully, what
running into it actually looks like: sshd accepting connections and
closing them before the banner, the web UI still responding normally,
clean ping, a dark panel, and a wrong clock after the next boot. None of
those read as "out of memory", which makes the failure hard to identify
from the symptoms.

Cross-referenced from SSH_UNAVAILABLE_AFTER_INSTALL.md, since "I can't
SSH in any more" is how most people will first meet this.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* fix: address review findings on the low-memory work

Nine CodeRabbit findings, five in code.

**Health state (the one that matters).** The non-dict guard did not cover a
dict missing fields the callers index directly, which is the shape actually
seen in the wild: a record carrying only circuit_state produced
`plugin clock-simple operation failed: 'circuit_state'` about fifty times a
minute with the panel frozen. The record is now completed against the
defaults per field rather than trusted or discarded wholesale. Per field
matters: a first pass rejected any incomplete record outright, which reset a
tripped breaker and real failure counts to healthy because one optional
field was absent -- an existing test caught it. Values of the wrong type
(a counter persisted as a string, an unknown circuit_state) fall back
individually, valid neighbours survive, and newer fields the schema has
grown since (degraded, degraded_reason) are carried through untouched.

**Cache ceiling.** MemoryCache.set() accepted entries without bound between
cleanup sweeps, which run every 300s by default, so a burst could take the
cache far past max_size -- the unbounded growth the limit exists to stop.
Eviction now runs under the same lock on every write, sharing one helper
with the periodic sweep so the two cannot drift.

**Installer, cgroups.** Only cgroup_enable=memory was checked, so a board
carrying that without cgroup_memory=1 reported success and got no change,
leaving MemoryMax= inert. Each parameter is now checked and appended
independently; verified against all four combinations, single line preserved.

**Installer, journald.** Persistence was inferred from /var/log/journal being
non-empty, which proves neither Storage=persistent nor a size cap -- the
directory survives a switch back to volatile. The effective configuration is
read instead (systemd-analyze cat-config, falling back to the conf files),
and an explicitly configured SystemMaxUse is preserved rather than
overwritten. Verified across volatile, persistent-without-cap,
persistent-with-user-cap, cap-without-storage, and commented-only configs.

**Dependency extras.** _extras_are_satisfied stopped at one level, so a
gated dependency that itself requests an extra (requests[socks]) passed on
the base distribution's version while the extra's own dependency was
missing, and pip was skipped. It now recurses, with a visited
(distribution, extras) set so a cycle terminates.

Docs: both kernel command-line paths documented (the installer falls back to
/boot/cmdline.txt), daemon-reload and restart added after the systemd
override example, memory exhaustion added to the SSH summary with its
power-cycle-only recovery, and a language on the fenced block for MD040.

Tests: five for the health-state repair including the exact wild shape and
that record_failure/record_success no longer raise against it, and one for
the cache ceiling. Both mutation-checked. Full suite 2927 passed, with the
one pre-existing tmpfs failure that also fails on main.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01STMbQE4YctTacQXfbYqKuW

* fix: harden the health-state repair and confirm journald took effect

Second review round; all three findings were valid and two were bugs in the
repair added last commit.

The repair could raise out of itself. An unhashable circuit_state (a list or
dict on disk) hit `value in {...}` and raised TypeError -- from the code
whose whole job is to stop a malformed record crashing the caller. It now
requires a str before the membership test.

bool is a subclass of int, so True passed the timestamp check and then
compared as 1.0: enough to expire a cooldown the instant the breaker opened,
while False would stop the elapsed check firing at all. Timestamps now
exclude bool explicitly.

The regression test for the original crash was seeded with a record that
*contained* circuit_state, so it passed against the old raw-return behaviour
too -- the counters are read with .get(), so circuit_state is the only field
whose absence used to raise. Reseeded to omit it, and it now fails against
raw-return as intended.

journald: drop-ins apply in lexical order, so a local file sorting after
ledmatrix-persistent.conf still wins and writing ours proves nothing. The
effective Storage is re-read afterwards and a warning naming the diagnostic
command is printed if persistence is still not active, rather than reporting
a success that was not verified.

Full suite 2934 passed, same single pre-existing tmpfs failure.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01STMbQE4YctTacQXfbYqKuW

---------

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
2026-08-19 12:28:22 -04:00

871 lines
36 KiB
Python

"""
Plugin Loader
Handles plugin module imports, dependency installation, and class instantiation.
Extracted from PluginManager to improve separation of concerns.
"""
import importlib
import importlib.metadata
import importlib.util
import json
import os
import sys
import subprocess
import threading
from pathlib import Path
from typing import Dict, Any, List, Optional, Tuple, Type
import logging
from packaging.requirements import InvalidRequirement, Requirement
from src.exceptions import PluginError
from src.logging_config import get_logger
def requirements_has_real_deps(requirements_file: str) -> bool:
"""
Check whether a requirements.txt actually specifies anything to install.
Plugins that ship all their dependencies with LEDMatrix core often keep a
requirements.txt where every line is commented out, for documentation
purposes only. Running pip against such a file still pays the full
subprocess/resolver cost for zero effect, so callers should skip the
install step entirely when this returns False.
"""
try:
with open(requirements_file, 'r', encoding='utf-8') as fh:
for line in fh:
line = line.strip()
if line and not line.startswith('#'):
return True
except OSError:
# Let the caller's own file handling report the error.
return True
return False
def _extra_dependencies(dist_name: str, extras) -> Optional[List[Requirement]]:
"""Dependencies a distribution declares *only* behind the given extras.
Returns None when the installed metadata cannot be read or parsed, so the
caller can fall back to running pip rather than assuming anything.
"""
try:
meta = importlib.metadata.metadata(dist_name)
except importlib.metadata.PackageNotFoundError:
return None
gated: List[Requirement] = []
for raw in meta.get_all('Requires-Dist') or []:
try:
dep = Requirement(raw)
except InvalidRequirement:
return None
if dep.marker is None:
continue
# Keep only what the distribution gates behind an extra we asked for:
# satisfied when `extra` is that name, but not when no extra is
# requested. A marker that holds either way (python_version, sys_platform)
# belongs to the base install and is already covered by the version check.
if dep.marker.evaluate({'extra': ''}):
continue
if any(dep.marker.evaluate({'extra': extra}) for extra in extras):
gated.append(dep)
return gated
def _extras_are_satisfied(req: Requirement, _visited: Optional[set] = None) -> bool:
"""Check the dependencies pulled in by req's extras are installed.
Follows extras through nested extras. A gated dependency can itself request
one (`requests[socks]`), and checking only that `requests` is installed at
an acceptable version says nothing about whether the socks extra's own
dependency is there -- so the caller would skip pip and the plugin would
fail at import instead. Plain dependencies are still checked one level
deep, which is all that is needed to tell "the extra was installed" from
"the extra was never installed".
`_visited` carries the (distribution, extras) pairs already seen, so a
dependency cycle between extras terminates instead of recursing forever.
Anything unreadable returns False, so the caller still falls through to pip.
"""
if _visited is None:
_visited = set()
marker = (req.name.lower(), frozenset(e.lower() for e in req.extras))
if marker in _visited:
# Already accounted for higher up the chain; treating a cycle as
# satisfied here is safe because the outer frame still has to pass.
return True
_visited.add(marker)
gated = _extra_dependencies(req.name, req.extras)
if gated is None:
return False
for dep in gated:
try:
dep_version = importlib.metadata.version(dep.name)
except importlib.metadata.PackageNotFoundError:
return False
if dep.specifier and not dep.specifier.contains(dep_version, prereleases=True):
return False
if dep.extras and not _extras_are_satisfied(dep, _visited):
return False
return True
def requirements_are_satisfied(requirements_file: str) -> bool:
"""
Check whether every real requirement line in requirements.txt is already
satisfied by packages installed in the current interpreter.
This replaces marker-file tracking with a direct fact check, so it's
immune to stale/missing/corrupted markers: it looks at what's actually
importable right now rather than trusting a hash comparison from a
previous run. Anything ambiguous (pip options, unparseable lines,
extras, unresolvable versions) conservatively returns False so the
caller falls through to running pip — this check only ever saves work,
never masks a real install.
"""
try:
with open(requirements_file, 'r', encoding='utf-8') as fh:
lines = fh.readlines()
except OSError:
return False
for raw_line in lines:
line = raw_line.strip()
if not line or line.startswith('#'):
continue
if line.startswith('-'):
return False # pip option (-r, --index-url, ...), can't verify
try:
req = Requirement(line)
except InvalidRequirement:
return False
if req.marker is not None and not req.marker.evaluate():
continue # not applicable on this platform/interpreter
try:
installed_version = importlib.metadata.version(req.name)
except importlib.metadata.PackageNotFoundError:
return False
if req.specifier and not req.specifier.contains(installed_version, prereleases=True):
return False
if req.extras and not _extras_are_satisfied(req):
return False
return True
def find_trusted_subdir(trusted_dir: str, name: str) -> Optional[str]:
"""Return `name` if it names an actual subdirectory of trusted_dir, else None.
Used as a containment check for a directory name derived from untrusted
input (a manifest-declared plugin id, an externally-supplied plugin
path): the returned value always comes from enumerating trusted_dir
itself via os.scandir(), so a caller that builds a path by joining
trusted_dir with this return value is joining against a name the
filesystem produced under a trusted root -- not the caller's original
string, which could otherwise smuggle a traversal sequence through.
"""
try:
with os.scandir(trusted_dir) as entries:
for entry in entries:
if entry.name == name and entry.is_dir():
return entry.name
except OSError:
pass
return None
class PluginLoader:
"""Handles plugin module loading and class instantiation."""
def __init__(self, logger: Optional[logging.Logger] = None) -> None:
"""
Initialize the plugin loader.
Args:
logger: Optional logger instance
"""
self.logger = logger or get_logger(__name__)
self._loaded_modules: Dict[str, Any] = {}
self._plugin_module_registry: Dict[str, set] = {} # Maps plugin_id to set of module names
# Lock to serialize module loading when plugins share module names
# (e.g., scroll_display.py, game_renderer.py across sport plugins).
# During exec_module, bare-name sub-modules temporarily appear in
# sys.modules; the lock prevents concurrent plugins from seeing each
# other's entries. After exec_module, _namespace_plugin_modules
# moves those bare names to namespaced keys (e.g.
# _plg_basketball_scoreboard_scroll_display) so they never collide.
self._module_load_lock = threading.Lock()
def find_plugin_directory(
self,
plugin_id: str,
plugins_dir: Path,
plugin_directories: Optional[Dict[str, Path]] = None
) -> Optional[Path]:
"""
Find the plugin directory for a given plugin ID.
Tries multiple strategies:
1. Use plugin_directories mapping if available
2. Direct path matching
3. Case-insensitive directory matching
4. Manifest-based search
Args:
plugin_id: Plugin identifier
plugins_dir: Base plugins directory
plugin_directories: Optional mapping of plugin_id to directory
Returns:
Path to plugin directory or None if not found
"""
# Sanitize plugin_id — os.path.basename is a CodeQL-recognized path sanitizer
plugin_id = os.path.basename(plugin_id or '')
if not plugin_id:
return None
# Strategy 1: Use mapping from discovery
if plugin_directories and plugin_id in plugin_directories:
plugin_dir = plugin_directories[plugin_id]
if plugin_dir.exists():
self.logger.debug("Using plugin directory from discovery mapping: %s", plugin_dir)
return plugin_dir
# Strategy 2: Direct paths — resolve and validate they stay within plugins_dir
plugins_dir_resolved = plugins_dir.resolve()
for _candidate_name in (plugin_id, f"ledmatrix-{plugin_id}"):
_candidate = (plugins_dir_resolved / _candidate_name).resolve()
try:
_candidate.relative_to(plugins_dir_resolved)
except ValueError:
continue
if _candidate.exists():
return _candidate
# Strategy 3: Case-insensitive search
normalized_id = plugin_id.lower()
for item in plugins_dir.iterdir():
if not item.is_dir():
continue
item_name = item.name
if item_name.lower() == normalized_id:
return item
if item_name.lower() == f"ledmatrix-{plugin_id}".lower():
return item
# Strategy 4: Manifest-based search
self.logger.debug("Directory name search failed for %s, searching by manifest...", plugin_id)
for item in plugins_dir.iterdir():
if not item.is_dir():
continue
# Skip if already checked
if item.name.lower() == normalized_id or item.name.lower() == f"ledmatrix-{plugin_id}".lower():
continue
manifest_path = item / "manifest.json"
if manifest_path.exists():
try:
with open(manifest_path, 'r', encoding='utf-8') as f:
item_manifest = json.load(f)
item_manifest_id = item_manifest.get('id')
if item_manifest_id == plugin_id:
self.logger.info(
"Found plugin %s in directory %s (manifest ID matches)",
plugin_id,
item.name
)
return item
except (json.JSONDecodeError, Exception) as e:
self.logger.debug("Skipping %s due to manifest error: %s", item.name, e)
continue
return None
def install_dependencies(
self,
plugin_dir: Path,
plugin_id: str,
plugins_dir: Path,
timeout: int = 300
) -> bool:
"""
Install plugin dependencies from requirements.txt.
Args:
plugin_dir: Plugin directory path
plugin_id: Plugin identifier
plugins_dir: Trusted base plugins directory for path containment check.
Required (not optional) so every caller reconstructs the plugin
path through the sanitiser below rather than trusting plugin_dir
directly -- CodeQL's path-injection query (and a malicious
manifest/plugin_id in practice) can't tell a legitimate
plugin_dir from one crafted to traverse outside plugins_dir.
timeout: Installation timeout in seconds
Returns:
True if dependencies installed or not needed, False on error
"""
plugin_id = os.path.basename(plugin_id or '')
if not plugin_id:
return False
# Resolve to a canonical absolute path (normalises .. and symlinks)
plugin_dir_real = os.path.realpath(str(plugin_dir))
plugins_dir_real = os.path.realpath(str(plugins_dir))
requested_name = os.path.basename(plugin_dir_real)
# Match the requested directory against an entry actually enumerated
# from the trusted plugins_dir, and build the path from that entry --
# not from requested_name. A name that came out of os.scandir() on a
# trusted root carries no taint regardless of what the caller asked
# for, so this is a real containment guarantee (an allowlist check
# against a trusted source), not a string-sanitisation of untrusted
# input that a static analyzer has to trust blindly.
matched_name = find_trusted_subdir(plugins_dir_real, requested_name)
if matched_name is None:
self.logger.error(
"Plugin directory for %s not found inside plugins dir", plugin_id
)
return False
safe_plugin_dir = os.path.join(plugins_dir_real, matched_name)
requirements_file = os.path.join(safe_plugin_dir, "requirements.txt")
if not os.path.isfile(requirements_file):
return True # No dependencies needed
if not requirements_has_real_deps(requirements_file):
self.logger.debug(
"requirements.txt for %s has no real dependencies (comments/blank only), skipping pip",
plugin_id
)
return True
if requirements_are_satisfied(requirements_file):
self.logger.debug(
"Dependencies for %s already satisfied in current environment, skipping pip",
plugin_id
)
return True
try:
self.logger.info("Installing dependencies for plugin %s...", plugin_id)
result = subprocess.run(
[sys.executable, "-m", "pip", "install", "--break-system-packages", "-r", requirements_file],
capture_output=True,
text=True,
timeout=timeout,
check=False
)
if result.returncode == 0:
self.logger.info("Dependencies installed successfully for %s", plugin_id)
return True
else:
stderr = result.stderr or ""
# uninstall-no-record-file means a system-managed copy of a package
# (e.g. apt's python3-requests, which ships no pip RECORD file) is in
# the way of the version this requirements.txt pins. Retry with
# --ignore-installed so pip lays the pinned version down alongside
# the system copy instead of trying to replace it — matching the
# retry already used by install_dependencies_apt.py / safe_pip_install.sh.
# Without this retry, the plugin would silently keep running against
# whatever version the system happened to ship.
if "uninstall-no-record-file" in stderr:
self.logger.warning(
"Dependencies for %s conflict with a system-managed package "
"(no pip RECORD); retrying with --ignore-installed: %s",
plugin_id, stderr.strip()
)
# Wrapped in its own try/except so a retry timeout is
# tolerated the same way as a retry failure, instead of
# propagating to the outer handler and returning False
# (which would contradict the "assume satisfied" fallback
# below).
try:
# sys.executable is this process's own interpreter (not
# attacker-influenced), and requirements_file is a path
# built internally by find_plugin_directory, never raw
# external input.
retry_result = subprocess.run( # nosec B603 - no shell invoked (list-form argv) # nosemgrep
[sys.executable, "-m", "pip", "install", "--break-system-packages",
"--ignore-installed", "-r", requirements_file],
capture_output=True,
text=True,
timeout=timeout,
check=False
)
if retry_result.returncode != 0:
self.logger.warning(
"Retry with --ignore-installed also failed for %s; assuming the "
"system-managed version satisfies the requirement: %s",
plugin_id, (retry_result.stderr or "").strip()
)
except subprocess.TimeoutExpired:
self.logger.warning(
"Retry with --ignore-installed timed out for %s; assuming the "
"system-managed version satisfies the requirement",
plugin_id
)
return True
self.logger.warning(
"Dependency installation returned non-zero exit code for %s: %s",
plugin_id,
stderr
)
return False
except subprocess.TimeoutExpired:
self.logger.error("Dependency installation timed out for %s", plugin_id)
return False
except FileNotFoundError:
self.logger.warning("pip not found. Skipping dependency installation for %s", plugin_id)
return True
except (BrokenPipeError, OSError) as e:
# Handle broken pipe errors (errno 32) which can occur during pip downloads
# Often caused by network interruptions or output buffer issues
if isinstance(e, OSError) and e.errno == 32:
self.logger.error(
"Broken pipe error during dependency installation for %s. "
"This usually indicates a network interruption or pip output buffer issue. "
"Try installing again or check your network connection.", plugin_id
)
else:
self.logger.error("OS error during dependency installation for %s: %s", plugin_id, e)
return False
except Exception as e:
self.logger.error("Unexpected error installing dependencies for %s: %s", plugin_id, e, exc_info=True)
return False
@staticmethod
def _iter_plugin_bare_modules(
plugin_dir: Path, before_keys: set
) -> list:
"""Return bare-name modules from plugin_dir added after before_keys.
Returns a list of (mod_name, module) tuples for modules that:
- Were added to sys.modules after before_keys snapshot
- Have bare names (no dots)
- Have a ``__file__`` inside plugin_dir
"""
resolved_dir = plugin_dir.resolve()
result = []
for key in set(sys.modules.keys()) - before_keys:
if "." in key:
continue
mod = sys.modules.get(key)
if mod is None:
continue
mod_file = getattr(mod, "__file__", None)
if not mod_file:
continue
try:
if Path(mod_file).resolve().is_relative_to(resolved_dir):
result.append((key, mod))
except (ValueError, TypeError):
continue
return result
def _evict_stale_bare_modules(self, plugin_dir: Path) -> dict:
"""Temporarily remove bare-name sys.modules entries from other plugins.
Before exec_module, scan the current plugin directory for .py files.
For each, if sys.modules has a bare-name entry whose ``__file__`` lives
in a *different* directory, remove it so Python's import system will
load the current plugin's version instead of reusing the stale cache.
Returns:
Dict mapping evicted module names to their module objects
(for restoration on error).
"""
resolved_dir = plugin_dir.resolve()
evicted: dict = {}
for py_file in plugin_dir.glob("*.py"):
mod_name = py_file.stem
if mod_name.startswith("_"):
continue
existing = sys.modules.get(mod_name)
if existing is None:
continue
existing_file = getattr(existing, "__file__", None)
if not existing_file:
continue
try:
if not Path(existing_file).resolve().is_relative_to(resolved_dir):
evicted[mod_name] = sys.modules.pop(mod_name)
self.logger.debug(
"Evicted stale bare-name module '%s' before loading plugin", mod_name,
)
except (ValueError, TypeError):
continue
return evicted
def _namespace_plugin_modules(
self, plugin_id: str, plugin_dir: Path, before_keys: set
) -> None:
"""
Move bare-name plugin modules to namespaced keys in sys.modules.
After exec_module loads a plugin's entry point, Python will have added
the plugin's local modules (scroll_display, game_renderer, …) to
sys.modules under their bare names. This method renames them to
``_plg_<plugin_id>_<module>`` so they cannot collide with identically-
named modules from other plugins.
The plugin code keeps working because ``from scroll_display import X``
binds ``X`` to the class *object*, not to the sys.modules entry.
Args:
plugin_id: Plugin identifier
plugin_dir: Plugin directory path
before_keys: Snapshot of sys.modules keys taken *before* exec_module
"""
safe_id = plugin_id.replace("-", "_")
namespaced_names: set = set()
for mod_name, mod in self._iter_plugin_bare_modules(plugin_dir, before_keys):
namespaced = f"_plg_{safe_id}_{mod_name}"
sys.modules[namespaced] = mod
# Remove the bare sys.modules entry. The module object stays
# alive via the namespaced key and all existing Python-level
# bindings (``from scroll_display import X`` already bound X
# to the class object). Leaving bare entries would cause the
# NEXT plugin's exec_module to find the cached entry and reuse
# it instead of loading its own version.
sys.modules.pop(mod_name, None)
namespaced_names.add(namespaced)
self.logger.debug(
"Namespace-isolated module '%s' -> '%s' for plugin %s",
mod_name, namespaced, plugin_id,
)
# Track for cleanup during unload
self._plugin_module_registry[plugin_id] = namespaced_names
if namespaced_names:
self.logger.info(
"Namespace-isolated %d module(s) for plugin %s",
len(namespaced_names), plugin_id,
)
def unregister_plugin_modules(self, plugin_id: str) -> None:
"""Remove namespaced sub-modules and cached module for a plugin from sys.modules.
Called by PluginManager during unload to clean up all module entries
that were created when the plugin was loaded.
"""
for ns_name in self._plugin_module_registry.pop(plugin_id, set()):
sys.modules.pop(ns_name, None)
self._loaded_modules.pop(plugin_id, None)
def load_module(
self,
plugin_id: str,
plugin_dir: Path,
entry_point: str
) -> Optional[Any]:
"""
Load a plugin module from file.
Module loading is serialized via _module_load_lock because plugins are
loaded in parallel (ThreadPoolExecutor) and multiple sport plugins
share identically-named local modules (scroll_display.py,
game_renderer.py, sports.py, etc.).
After loading, bare-name modules from the plugin directory are moved
to namespaced keys in sys.modules (e.g. ``_plg_basketball_scoreboard_scroll_display``)
so they cannot collide with other plugins.
Args:
plugin_id: Plugin identifier
plugin_dir: Plugin directory path
entry_point: Entry point filename (e.g., 'manager.py')
Returns:
Loaded module or None on error
"""
plugin_id = os.path.basename(plugin_id or '')
if not plugin_id:
raise PluginError("Invalid plugin ID")
try:
plugin_dir_resolved = plugin_dir.resolve(strict=True)
except OSError:
raise PluginError("Plugin directory not found", plugin_id=plugin_id)
entry_file = (plugin_dir_resolved / entry_point).resolve()
try:
entry_file.relative_to(plugin_dir_resolved)
except ValueError:
raise PluginError("Invalid entry point path", plugin_id=plugin_id)
if not entry_file.exists():
error_msg = f"Entry point file not found for plugin {plugin_id}"
self.logger.error(error_msg)
raise PluginError(error_msg, plugin_id=plugin_id, context={'entry_file': str(entry_file)})
with self._module_load_lock:
# Add plugin directory to sys.path if not already there
plugin_dir_str = str(plugin_dir)
if plugin_dir_str not in sys.path:
sys.path.insert(0, plugin_dir_str)
self.logger.debug("Added plugin %s's directory to sys.path", plugin_id)
# Import the plugin module
module_name = f"plugin_{plugin_id.replace('-', '_')}"
# Check if already loaded
if module_name in sys.modules:
self.logger.debug("Module %s already loaded, reusing", module_name)
return sys.modules[module_name]
spec = importlib.util.spec_from_file_location(module_name, entry_file)
if spec is None or spec.loader is None:
self.logger.error("Could not create module spec for plugin %s", plugin_id)
error_msg = f"Could not create module spec for {entry_file}"
raise PluginError(error_msg, plugin_id=plugin_id, context={'entry_file': str(entry_file)})
module = importlib.util.module_from_spec(spec)
sys.modules[module_name] = module
# Snapshot AFTER inserting the main module so that
# _namespace_plugin_modules and error cleanup only target
# sub-modules, not the main module entry itself.
before_keys = set(sys.modules.keys())
# Evict stale bare-name modules from other plugin directories
# so Python's import system loads fresh copies from this plugin.
evicted = self._evict_stale_bare_modules(plugin_dir)
try:
spec.loader.exec_module(module)
# Move bare-name plugin modules to namespaced keys so they
# cannot collide with identically-named modules from other plugins
self._namespace_plugin_modules(plugin_id, plugin_dir, before_keys)
except Exception:
# Restore evicted modules so other plugins are unaffected
for evicted_name, evicted_mod in evicted.items():
if evicted_name not in sys.modules:
sys.modules[evicted_name] = evicted_mod
# Clean up the partially-initialized main module and any
# bare-name sub-modules that were added during exec_module
# so they don't leak into subsequent plugin loads.
sys.modules.pop(module_name, None)
for key, _ in self._iter_plugin_bare_modules(plugin_dir, before_keys):
sys.modules.pop(key, None)
raise
self._loaded_modules[plugin_id] = module
self.logger.debug("Loaded module %s for plugin %s", module_name, plugin_id)
return module
def get_plugin_class(
self,
plugin_id: str,
module: Any,
class_name: str
) -> Type[Any]:
"""
Get the plugin class from a loaded module.
Args:
plugin_id: Plugin identifier
module: Loaded module
class_name: Name of the plugin class
Returns:
Plugin class
Raises:
PluginError: If class not found
"""
if not hasattr(module, class_name):
error_msg = f"Class {class_name} not found in module for plugin {plugin_id}"
self.logger.error(error_msg)
raise PluginError(
error_msg,
plugin_id=plugin_id,
context={'class_name': class_name, 'module': module.__name__}
)
plugin_class = getattr(module, class_name)
# Verify it's a class
if not isinstance(plugin_class, type):
error_msg = f"{class_name} is not a class in module for plugin {plugin_id}"
self.logger.error(error_msg)
raise PluginError(error_msg, plugin_id=plugin_id, context={'class_name': class_name})
return plugin_class
def instantiate_plugin(
self,
plugin_id: str,
plugin_class: Type[Any],
config: Dict[str, Any],
display_manager: Any,
cache_manager: Any,
plugin_manager: Any
) -> Any:
"""
Instantiate a plugin class.
Args:
plugin_id: Plugin identifier
plugin_class: Plugin class to instantiate
config: Plugin configuration
display_manager: Display manager instance
cache_manager: Cache manager instance
plugin_manager: Plugin manager instance
Returns:
Plugin instance
Raises:
PluginError: If instantiation fails
"""
try:
plugin_instance = plugin_class(
plugin_id=plugin_id,
config=config,
display_manager=display_manager,
cache_manager=cache_manager,
plugin_manager=plugin_manager
)
self.logger.debug("Instantiated plugin %s", plugin_id)
return plugin_instance
except Exception as e:
error_msg = f"Failed to instantiate plugin {plugin_id}: {e}"
self.logger.error(error_msg, exc_info=True)
raise PluginError(error_msg, plugin_id=plugin_id) from e
@staticmethod
def _parse_semver(value: Any) -> Optional[Tuple[int, int, int]]:
"""Parse 'X.Y.Z' (extra parts/suffixes ignored) into a comparable
3-tuple, or None when unparseable."""
if not isinstance(value, str):
return None
parts = value.strip().lstrip('v').split('.')
try:
nums = [int(''.join(ch for ch in p if ch.isdigit()) or 0) for p in parts[:3]]
except ValueError:
return None
while len(nums) < 3:
nums.append(0)
return tuple(nums) # type: ignore[return-value]
def _warn_if_incompatible(self, plugin_id: str, manifest: Dict[str, Any]) -> None:
"""Log one warning when a plugin declares a minimum LEDMatrix version
newer than the running core. Advisory only — never raises — so a
plugin that guards optional features with try/except keeps working.
"""
from src import __version__ as core_version
from src.plugin_system import compatibility
compatible, _reason = compatibility.check(manifest, core_version)
if compatible:
# Distinguish "fine" from "couldn't tell" for anyone reading logs:
# a core below the trustworthy floor is skipped, not cleared.
current = compatibility.parse_semver(core_version)
if current is None or current < compatibility.TRUSTWORTHY_FLOOR:
self.logger.debug(
"Skipping version compatibility check for %s: core __version__ "
"(%s) is below the ecosystem floor", plugin_id, core_version)
return
declared = compatibility.declared_min_version(manifest)
self.logger.warning(
"Plugin %s declares min LEDMatrix version %s but this core is %s — "
"features it relies on may be missing; update the core or expect "
"degraded fallbacks", plugin_id, declared, core_version)
def load_plugin(
self,
plugin_id: str,
manifest: Dict[str, Any],
plugin_dir: Path,
config: Dict[str, Any],
display_manager: Any,
cache_manager: Any,
plugin_manager: Any,
install_deps: bool = True,
plugins_dir: Optional[Path] = None,
) -> Tuple[Any, Any]:
"""
Complete plugin loading process.
Args:
plugin_id: Plugin identifier
manifest: Plugin manifest
plugin_dir: Plugin directory path
config: Plugin configuration
display_manager: Display manager instance
cache_manager: Cache manager instance
plugin_manager: Plugin manager instance
install_deps: Whether to install dependencies
plugins_dir: Trusted base plugins directory forwarded to install_dependencies
Returns:
Tuple of (plugin_instance, module)
Raises:
PluginError: If loading fails
"""
self._warn_if_incompatible(plugin_id, manifest)
# Install dependencies if needed
if install_deps:
if plugins_dir is None:
raise PluginError(
f"plugins_dir is required to install dependencies for plugin {plugin_id} "
"(needed for path containment; pass install_deps=False if the caller "
"doesn't have a trusted plugins directory to supply)",
plugin_id=plugin_id,
context={'plugin_dir': str(plugin_dir)},
)
if not self.install_dependencies(plugin_dir, plugin_id, plugins_dir=plugins_dir):
raise PluginError(
f"Dependency installation failed for plugin {plugin_id} in {plugin_dir}",
plugin_id=plugin_id,
context={'plugin_dir': str(plugin_dir)},
)
# Load module
entry_point = manifest.get('entry_point', 'manager.py')
module = self.load_module(plugin_id, plugin_dir, entry_point)
if module is None:
raise PluginError(f"Failed to load module for plugin {plugin_id}", plugin_id=plugin_id)
# Get plugin class
class_name = manifest.get('class_name')
if not class_name:
raise PluginError(f"No class_name in manifest for plugin {plugin_id}", plugin_id=plugin_id)
plugin_class = self.get_plugin_class(plugin_id, module, class_name)
# Instantiate plugin
plugin_instance = self.instantiate_plugin(
plugin_id,
plugin_class,
config,
display_manager,
cache_manager,
plugin_manager
)
return (plugin_instance, module)