MRG: add plugin support for new command-line subcommands

doc/command-line.md

@@ -1966,3 +1966,19 @@ situations where you have a **very large** collection of signatures
 in the collection (as you would have to, with a zipfile). This can be
 useful if you want to refer to different subsets of the collection
 without making multiple copies in a zip file.
+
+### Using sourmash plugins
+
+As of sourmash v4.7.0, sourmash has an experimental plugins interface!
+The plugin interface supports extending sourmash to load and save
+signatures in new ways, and also supports the addition of sourmash
+subcommands via `sourmash scripts`.
+
+In order to use a plugin with sourmash, you will need to use `pip`
+or `conda` to install the plugin the same environment that sourmash
+is installed in.
+
+In the future, we will include a list of available sourmash plugins in
+the documentation, and also provide a way to list available plugins.
+
+You can list all installed plugins with `sourmash info -v`.

doc/dev_plugins.md

@@ -2,7 +2,8 @@
 
 As of version 4.7.0, sourmash has experimental support for Python
 plugins to load and save signatures in different ways (e.g. file
-formats, RPC servers, databases, etc.).  This support is provided via
+formats, RPC servers, databases, etc.) and to run additional commands
+via the command-line.  This support is provided via
 the "entry points" mechanism supplied by
 [`importlib.metadata`](https://docs.python.org/3/library/importlib.metadata.html)
 and documented
@@ -24,23 +25,38 @@ a_reader = "module_name:load_sketches"
 
 [project.entry-points."sourmash.save_to"]
 a_writer = "module_name:SaveSignatures_WriteFile"
+
+[project.entry-points."sourmash.cli_script"]
+new_cli = "module_name:Command_NewCommand"
 ```
 
 Here, `module_name` should be the name of the module to import.
-`load_sketches` should be a function that takes a location along with
+
+* `load_sketches` should be a function that takes a location along with
 arbitrary keyword arguments and returns an `Index` object
 (e.g. `LinearIndex` for a collection of in-memory
-signatures). `SaveSignatures_WriteFile` should be a class that
+signatures).
+* `SaveSignatures_WriteFile` should be a class that
 subclasses `BaseSave_SignaturesToLocation` and implements its own
 mechanisms of saving signatures. See the `sourmash.save_load` module
 for saving and loading code already used in sourmash.
-
-Note that if the function or class has a `priority` attribute, this will
-be used to determine the order in which the plugins are called.
-
-The `name` attribute of the plugin (`a_reader` and `a_writer` in
+* `Command_NewCommand` should be a class that subclasses
+  `plugins.CommandLinePlugin` and provides an `__init__` and
+  `main` method.
+
+Note that if the reader function or writer class has a `priority`
+attribute, this will be used to determine the order in which the
+plugins are called. Priorities lower than 10 will get called before
+any internal load or save function, while priorities greater than 80
+will get called after almost all internal load/save functions; see
+`src/sourmash/save_load.py` for details and the current priorities.
+
+The `name` attribute of the plugin (`a_reader`, `a_writer`, and `new_cli` in
 `pyproject.toml`, above) is only used in debugging.
 
+You can provide zero or more plugins, and you can define just a reader, or
+just a writer, or just a CLI plugin.
+
 ## Templates and examples
 
 If you want to create your own plug-in, you can start with the
@@ -53,15 +69,19 @@ Some (early stage) plugins are also available as examples:
 
 ## Debugging plugins
 
+`sourmash info -v` will list all installed plugins.
+
 `sourmash sig cat <input sig> -o <output sig>` is a simple way to
 invoke a `save_to` plugin. Use `-d` to turn on debugging output.
 
 `sourmash sig describe <input location>` is a simple way to invoke
 a `load_from` plugin. Use `-d` to turn on debugging output.
 
+`sourmash scripts` will list available command-line plugins.
+
 ## Semantic versioning and listing sourmash as a dependency
 
-Plugins should probably list sourmash as a dependency for installation.
+Plugins should generally list sourmash as a dependency for installation.
 
 Once plugins are officially supported by sourmash, the plugin API will
 be under [semantic versioning constraints](https://semver.org/). That

src/sourmash/__main__.py

@@ -1,7 +1,12 @@
-import sourmash
+"""
+The main entry point for sourmash, defined in pyproject.toml.
+
+Can also be executed as 'python -m sourmash'.
+"""
 
 
 def main(arglist=None):
+    import sourmash
     args = sourmash.cli.get_parser().parse_args(arglist)
     if hasattr(args, 'subcmd'):
         mod = getattr(sourmash.cli, args.cmd)
@@ -10,7 +15,9 @@ def main(arglist=None):
     else:
         mod = getattr(sourmash.cli, args.cmd)
         mainmethod = getattr(mod, 'main')
-    return mainmethod(args)
+
+    retval = mainmethod(args)
+    raise SystemExit(retval)
 
 
 if __name__ == '__main__':

src/sourmash/cli/__init__.py

@@ -38,6 +38,7 @@
 from . import sketch
 from . import storage
 from . import tax
+from . import scripts
 
 
 class SourmashParser(ArgumentParser):
@@ -98,20 +99,25 @@ def get_parser():
         'sketch': 'Create signatures',
         'sig': 'Manipulate signature files',
         'storage': 'Operations on storage',
+        'scripts': "Plug-ins",
     }
     alias = {
-        "sig": "signature"
+        "sig": "signature",
+        "ext": "scripts",
     }
     expert = set(['categorize', 'import_csv', 'migrate', 'multigather', 'sbt_combine', 'watch'])
 
     clidir = os.path.dirname(__file__)
     basic_ops = utils.command_list(clidir)
-    user_ops = [op for op in basic_ops if op not in expert]
+
+    # provide a list of the basic operations - not expert, not submodules.
+    user_ops = [op for op in basic_ops if op not in expert and op not in module_descs]
     usage = '    Basic operations\n'
     for op in user_ops:
         docstring = getattr(sys.modules[__name__], op).__doc__
         helpstring = 'sourmash {op:s} --help'.format(op=op)
         usage += '        {hs:25s} {ds:s}\n'.format(hs=helpstring, ds=docstring)
+    # next, all the subcommand ones - dive into subdirectories.
     cmd_group_dirs = next(os.walk(clidir))[1]
     cmd_group_dirs = filter(utils.opfilter, cmd_group_dirs)
     cmd_group_dirs = sorted(cmd_group_dirs)

src/sourmash/cli/info.py

@@ -4,6 +4,7 @@
 import screed
 import sourmash
 from sourmash.logging import notify
+from sourmash.plugins import list_all_plugins
 
 def subparser(subparsers):
     subparser = subparsers.add_parser('info')
@@ -26,6 +27,8 @@ def info(verbose=False):
         notify(f'screed version {screed.__version__}')
         notify(f'- loaded from path: {os.path.dirname(screed.__file__)}')
 
+        list_all_plugins()
+
 
 def main(args):
     info(verbose=args.verbose)

src/sourmash/cli/scripts/__init__.py

@@ -0,0 +1,48 @@
+"""Provide a mechanism to add CLI plugins to sourmash.
+
+See https://sourmash.readthedocs.io/en/latest/dev_plugins.html for docs,
+src/sourmash/plugins.py for core sourmash implementation code, and
+https://github.com/sourmash-bio/sourmash_plugin_template for a template repo
+for making new plugins.
+"""
+
+# CTB TODO:
+# * provide suggestions for documentation & metadata for authors:
+# * provide guidance on how to test your CLI plugin at the CLI
+#   (minimal testing regime: sourmash scripts, look for description etc.)
+
+import argparse
+import sourmash
+
+# Here, we decorate this module with the various extension objects
+# e.g. 'sourmash scripts foo' will look up attribute 'scripts.foo'
+# and we will return the extension class object, which will then
+# be run by sourmash.__main__. This dictionary is loaded below
+# by sourmash.plugins.add_cli_scripts.
+_extension_dict = {}
+
+def __getattr__(name):
+    if name in _extension_dict:
+        return _extension_dict[name]
+    raise AttributeError(name)
+
+def subparser(subparsers):
+    subparser = subparsers.add_parser('scripts',
+                                      usage=argparse.SUPPRESS,
+                                      formatter_class=argparse.RawDescriptionHelpFormatter,
+                                      aliases=['ext'])
+
+    # get individual help strings:
+    descrs = list(sourmash.plugins.get_cli_scripts_descriptions())
+    if descrs:
+        description = "\n".join(descrs)
+    else:
+        description = "(No script plugins detected!)"
+
+    s = subparser.add_subparsers(title="available plugin/extension commands",
+                                 dest='subcmd',
+                                 metavar='subcmd',
+                                 help=argparse.SUPPRESS,
+                                 description=description)
+
+    _extension_dict.update(sourmash.plugins.add_cli_scripts(s))

src/sourmash/plugins.py

@@ -4,7 +4,7 @@
 Plugin entry point names:
 * 'sourmash.load_from' - Index class loading.
 * 'sourmash.save_to' - Signature saving.
-* 'sourmash.picklist_filters' - extended Picklist functionality.
+* 'sourmash.cli_script' - command-line extension.
 
 CTB TODO:
 
@@ -15,7 +15,9 @@
 DEFAULT_LOAD_FROM_PRIORITY = 99
 DEFAULT_SAVE_TO_PRIORITY = 99
 
-from .logging import debug_literal
+import itertools
+
+from .logging import (debug_literal, error, notify, set_quiet)
 
 # cover for older versions of Python that don't support selection on load
 # (the 'group=' below).
@@ -31,14 +33,23 @@
 # load 'save_to' entry points as well.
 _plugin_save_to = entry_points(group='sourmash.save_to')
 
+# aaaaand CLI entry points:
+_plugin_cli = entry_points(group='sourmash.cli_script')
+_plugin_cli_once = False
+
+###
 
 def get_load_from_functions():
     "Load the 'load_from' plugins and yield tuples (priority, name, fn)."
     debug_literal(f"load_from plugins: {_plugin_load_from}")
 
     # Load each plugin,
     for plugin in _plugin_load_from:
-        loader_fn = plugin.load()
+        try:
+            loader_fn = plugin.load()
+        except (ModuleNotFoundError, AttributeError) as e:
+            debug_literal(f"plugins.load_from_functions: got error loading {plugin.name}: {str(e)}")
+            continue
 
         # get 'priority' if it is available
         priority = getattr(loader_fn, 'priority', DEFAULT_LOAD_FROM_PRIORITY)
@@ -55,7 +66,11 @@ def get_save_to_functions():
 
     # Load each plugin,
     for plugin in _plugin_save_to:
-        save_cls = plugin.load()
+        try:
+            save_cls = plugin.load()
+        except (ModuleNotFoundError, AttributeError) as e:
+            debug_literal(f"plugins.load_from_functions: got error loading {plugin.name}: {str(e)}")
+            continue
 
         # get 'priority' if it is available
         priority = getattr(save_cls, 'priority', DEFAULT_SAVE_TO_PRIORITY)
@@ -64,3 +79,104 @@ def get_save_to_functions():
         name = plugin.name
         debug_literal(f"plugins.save_to_functions: got '{name}', priority={priority}")
         yield priority, save_cls
+
+
+class CommandLinePlugin:
+    """
+    Provide some minimal common CLI functionality - -q and -d.
+
+    Subclasses should call super().__init__(parser) and super().main(args).
+    """
+    command = None
+    description = None
+
+    def __init__(self, parser):
+        parser.add_argument(
+            '-q', '--quiet', action='store_true',
+            help='suppress non-error output'
+        )
+        parser.add_argument(
+            '-d', '--debug', action='store_true',
+            help='provide debugging output'
+        )
+
+    def main(self, args):
+        set_quiet(args.quiet, args.debug)
+
+
+def get_cli_script_plugins():
+    global _plugin_cli_once
+
+    x = []
+    for plugin in _plugin_cli:
+        name = plugin.name
+        mod = plugin.module
+        try:
+            script_cls = plugin.load()
+        except (ModuleNotFoundError, AttributeError):
+            if _plugin_cli_once is False:
+                error(f"ERROR: cannot find or load module for cli_script plugin '{name}'")
+            continue
+
+        command = getattr(script_cls, 'command', None)
+        if command is None:
+            # print error message only once...
+            if _plugin_cli_once is False:
+                error(f"ERROR: no command provided by cli_script plugin '{name}' from {mod}; skipping")
+        else:
+            x.append(plugin)
+
+    _plugin_cli_once = True
+    return x
+
+
+def get_cli_scripts_descriptions():
+    "Build the descriptions for command-line plugins."
+    for plugin in get_cli_script_plugins():
+        name = plugin.name
+        script_cls = plugin.load()
+
+        command = getattr(script_cls, 'command')
+        description = getattr(script_cls, 'description',
+                              f"(no description provided by plugin '{name}')")
+        yield f"sourmash scripts {command:16s} - {description}"
+
+
+def add_cli_scripts(parser):
+    "Configure parsing for command-line plugins."
+    d = {}
+
+    for plugin in get_cli_script_plugins():
+        name = plugin.name
+        script_cls = plugin.load()
+
+        subparser = parser.add_parser(script_cls.command)
+        debug_literal(f"cls_script plugin '{name}' adding command '{script_cls.command}'")
+        obj = script_cls(subparser)
+        d[script_cls.command] = obj
+
+    return d
+
+
+def list_all_plugins():
+    plugins = itertools.chain(_plugin_load_from,
+                              _plugin_save_to,
+                              _plugin_cli)
+    plugins = list(plugins)
+
+    if not plugins:
+        notify("\n(no plugins detected)\n")
+
+    notify("")
+    notify("the following plugins are installed:")
+    notify("")
+    notify(f"{'plugin type':<20s} {'from python module':<30s} {'v':<5s} {'entry point name':<20s}")
+    notify(f"{'-'*20} {'-'*30} {'-'*5} {'-'*20}")
+
+    for plugin in plugins:
+        name = plugin.name
+        mod = plugin.module
+        version = plugin.dist.version
+        group = plugin.group
+
+        notify(f"{group:<20s} {mod:<30s} {version:<5s} {name:<20s}")