Overhauled CLI. #180

- Everything now accepts -i, -e, -q, and -v.

- Plumbing commands now announce what database/index they're using and
  where they got them from.

- New command status, which does nothing but still does the announcing.

- New command reindex, which recreates only the whoosh index.

pokedex/__init__.py

@@ -1,8 +1,10 @@
 # encoding: utf8
 from optparse import OptionParser
+import os
+import pkg_resources
 import sys
 
-from .db import connect, metadata
+import pokedex.db
 import pokedex.db.load
 import pokedex.lookup
 
@@ -26,39 +28,178 @@ def main():
         command_help()
 
 
-def command_dump(*args):
+def get_parser(verbose=True):
-    parser = OptionParser()
+    """Returns an OptionParser prepopulated with the global options.
-    parser.add_option('-e', '--engine', dest='engine_uri', default=None)
+    
-    parser.add_option('-d', '--directory', dest='directory', default=None)
+    `verbose` is whether or not the options should be verbose by default.
-    parser.add_option('-q', '--quiet', dest='verbose', default=True, action='store_false')
+    """
-    options, _ = parser.parse_args(list(args))
+    parser = OptionParser()
+    parser.add_option('-e', '--engine', dest='engine_uri', default=os.environ.get('POKEDEX_DB_ENGINE', None))
+    parser.add_option('-i', '--index', dest='index_dir', default=os.environ.get('POKEDEX_INDEX_DIR', None))
+    parser.add_option('-q', '--quiet', dest='verbose', default=verbose, action='store_false')
+    parser.add_option('-v', '--verbose', dest='verbose', default=verbose, action='store_true')
+    return parser
+
+def get_session(options):
+    """Given a parsed options object, connects to the database and returns a
+    session.
+    """
+
+    engine_uri = options.engine_uri
+    got_from = None
+    if engine_uri:
+        got_from = 'command line'
+    else:
+        engine_uri = os.environ.get('POKEDEX_DB_ENGINE', None)
+        if engine_uri:
+            got_from = 'environment'
+        else:
+            got_from = 'default setting'
+
+    session = pokedex.db.connect(engine_uri)
+
+    if options.verbose:
+        print "Connected to database {engine} (from {got_from})" \
+            .format(engine=session.bind.url, got_from=got_from)
+
+    return session
+
+def get_lookup(options, session=None, recreate=False):
+    """Given a parsed options object, opens the whoosh index and returns a
+    PokedexLookup object.
+
+    Unlike `get_session`, this function can actually do population as a side
+    effect!  This is fallout from how PokedexLookup works.
+    """
+    # TODO fix the above
+
+    if recreate and not session:
+        raise ValueError("get_lookup() needs an explicit session to regen the index")
+
+    index_dir = options.index_dir
+    got_from = None
+    if index_dir:
+        got_from = 'command line'
+    else:
+        index_dir = os.environ.get('POKEDEX_INDEX_DIR', None)
+        if index_dir:
+            got_from = 'environment'
+        else:
+            index_dir = pkg_resources.resource_filename('pokedex',
+                                                        'data/whoosh-index')
+            got_from = 'default setting'
+
+    if options.verbose:
+        print "Opened lookup index {index_dir} (from {got_from})" \
+            .format(index_dir=index_dir, got_from=got_from)
+
+    lookup = pokedex.lookup.PokedexLookup(index_dir, session=session,
+                                                     recreate=recreate)
+
+    return lookup
+
+def print_csv_directory(options):
+    """Just prints the csv directory we're about to use."""
+
+    if not options.verbose:
+        return
+
+    if options.directory:
+        csvdir = options.directory
+        got_from = 'command line'
+    else:
+        # This is the same as the db.load default
+        csvdir = pkg_resources.resource_filename('pokedex', 'data/csv')
+        got_from = 'default setting'
+
+    print "Using CSV directory {csvdir} (from {got_from})" \
+        .format(csvdir=csvdir, got_from=got_from)
+
+
+### Plumbing commands
+
+def command_dump(*args):
+    parser = get_parser(verbose=True)
+    parser.add_option('-d', '--directory', dest='directory', default=None)
+    options, tables = parser.parse_args(list(args))
+
+    session = get_session(options)
+    print_csv_directory(options)
 
-    session = connect(options.engine_uri)
     pokedex.db.load.dump(session, directory=options.directory,
+                                  tables=tables,
                                   verbose=options.verbose)
 
 def command_load(*args):
-    parser = OptionParser()
+    parser = get_parser(verbose=True)
-    parser.add_option('-e', '--engine', dest='engine_uri', default=None)
     parser.add_option('-d', '--directory', dest='directory', default=None)
     parser.add_option('-D', '--drop-tables', dest='drop_tables', default=False, action='store_true')
-    parser.add_option('-q', '--quiet', dest='verbose', default=True, action='store_false')
+    options, tables = parser.parse_args(list(args))
-    options, _ = parser.parse_args(list(args))
 
-    session = connect(options.engine_uri)
+    if not options.engine_uri:
+        print "WARNING: You're reloading the default database, but not the lookup index.  They"
+        print "         might get out of sync, and pokedex commands may not work correctly!"
+        print "To fix this, run `pokedex reindex` when this command finishes.  Or, just use"
+        print "`pokedex setup` to do both at once."
+        print
+
+    session = get_session(options)
+    print_csv_directory(options)
 
     pokedex.db.load.load(session, directory=options.directory,
                                   drop_tables=options.drop_tables,
+                                  tables=tables,
                                   verbose=options.verbose)
 
+def command_reindex(*args):
+    parser = get_parser(verbose=True)
+    options, _ = parser.parse_args(list(args))
+
+    session = get_session(options)
+    lookup = get_lookup(options, session=session, recreate=True)
+
+    print "Recreated lookup index."
+
+
 def command_setup(*args):
-    session = connect()
+    parser = get_parser(verbose=False)
-    pokedex.db.load.load(session, verbose=False, drop_tables=True)
+    options, _ = parser.parse_args(list(args))
-    pokedex.lookup.PokedexLookup(session=session, recreate=True)
+
+    options.directory = None
+
+    session = get_session(options)
+    print_csv_directory(options)
+    pokedex.db.load.load(session, directory=None, drop_tables=True,
+                                  verbose=options.verbose)
+
+    lookup = get_lookup(options, session=session, recreate=True)
+
+    print "Recreated lookup index."
 
 
-def command_lookup(name):
+def command_status(*args):
-    results = pokedex.lookup.PokedexLookup().lookup(name)
+    parser = get_parser(verbose=True)
+    options, _ = parser.parse_args(list(args))
+    options.verbose = True
+    options.directory = None
+
+    session = get_session(options)
+    print_csv_directory(options)
+    lookup = get_lookup(options, recreate=False)
+
+
+### User-facing commands
+
+def command_lookup(*args):
+    parser = get_parser(verbose=False)
+    options, words = parser.parse_args(list(args))
+
+    name = u' '.join(words)
+
+    session = get_session(options)
+    lookup = get_lookup(options, session=session, recreate=False)
+
+    results = lookup.lookup(name)
     if not results:
         print "No matches."
     elif results[0].exact:
@@ -83,6 +224,7 @@ def command_help():
     print u"""pokedex -- a command-line Pokédex interface
 usage: pokedex {command} [options...]
 Run `pokedex setup` first, or nothing will work!
+See http://bugs.veekun.com/projects/pokedex/wiki/CLI for more documentation.
 
 Commands:
     help                Displays this message.
@@ -91,19 +233,33 @@ Commands:
 System commands:
     load                Load Pokédex data into a database from CSV files.
     dump                Dump Pokédex data from a database into CSV files.
-    setup               Loads Pokédex data into the right place and creates a
+    reindex             Rebuilds the lookup index from the database.
-                        lookup index in the right place.  No options or output.
+    setup               Combines load and reindex.
-                        This will blow away the default database and index!
+    status              No effect, but prints which engine, index, and csv
+                        directory would be used for other commands.
 
-Options:
+Global options:
+    -e|--engine=URI     By default, all commands try to use a SQLite database
+                        in the pokedex install directory.  Use this option (or
+                        a POKEDEX_DB_ENGINE environment variable) to specify an
+                        alternate database.
+    -i|--index=DIR      By default, all commands try to put the lookup index in
+                        the pokedex install directory.  Use this option (or a
+                        POKEDEX_INDEX_DIR environment variable) to specify an
+                        alternate loction.
+
+System options:
     -d|--directory      By default, load and dump will use the CSV files in the
                         pokedex install directory.  Use this option to specify
                         a different directory.
     -D|--drop-tables    With load, drop all tables before loading data.
-    -e|--engine=URI     By default, all commands try to use a SQLite database
+    -q|--quiet          Don't print system output.  This is the default for
-                        in the pokedex install directory.  Use this option to
+                        non-system commands and setup.
-                        specify an alternate database.
+    -v|--verbose        Print system output.  This is the default for system
-    -q|--quiet          Turn off any unnecessary status output from dump/load.
+                        commands, except setup.
+
+    Additionally, load and dump accept a list of table names (possibly with
+    wildcards) and/or csv fileames as an argument list.
 """.encode(sys.getdefaultencoding(), 'replace')
 
     sys.exit(0)

pokedex/db/load.py

@@ -1,15 +1,49 @@
 """CSV to database or vice versa."""
 import csv
+import os.path
 import pkg_resources
+import re
 import sys
 
 from sqlalchemy.orm.attributes import instrumentation_registry
+import sqlalchemy.sql.util
 import sqlalchemy.types
 
 from pokedex.db import metadata
 import pokedex.db.tables as tables
 
 
+def _wildcard_char_to_regex(char):
+    """Converts a single wildcard character to the regex equivalent."""
+
+    if char == '?':
+        return '.?'
+    elif char == '*':
+        return '.*'
+    else:
+        return re.escape(char)
+
+def _wildcard_glob_to_regex(glob):
+    """Converts a single wildcard glob to a regex STRING."""
+
+    # If it looks like a filename, make it not one
+    if '.' in glob or '/' in glob:
+        _, filename = os.path.split(glob)
+        table_name, _ = os.path.splitext(filename)
+        glob = table_name
+
+    return u''.join(map(_wildcard_char_to_regex, glob))
+
+def _wildcards_to_regex(strings):
+    """Converts a list of wildcard globs to a single regex object."""
+
+    regex_parts = map(_wildcard_glob_to_regex, strings)
+
+    regex = '^(?:' + '|'.join(regex_parts) + ')$'
+
+    return re.compile(regex)
+
+
 def _get_verbose_prints(verbose):
     """If `verbose` is true, returns two functions: one for printing a starting
     message, and the other for printing a success or failure message when
@@ -44,7 +78,7 @@ def _get_verbose_prints(verbose):
     return dummy, dummy
 
 
-def load(session, directory=None, drop_tables=False, verbose=False):
+def load(session, tables=[], directory=None, drop_tables=False, verbose=False):
     """Load data from CSV files into the given database session.
 
     Tables are created automatically.
@@ -52,6 +86,9 @@ def load(session, directory=None, drop_tables=False, verbose=False):
     `session`
         SQLAlchemy session to use.
 
+    `tables`
+        List of tables to load.  If omitted, all tables are loaded.
+
     `directory`
         Directory the CSV files reside in.  Defaults to the `pokedex` data
         directory.
@@ -70,17 +107,29 @@ def load(session, directory=None, drop_tables=False, verbose=False):
     if not directory:
         directory = pkg_resources.resource_filename('pokedex', 'data/csv')
 
+    if tables:
+        regex = _wildcards_to_regex(tables)
+        table_names = filter(regex.match, metadata.tables.keys())
+    else:
+        table_names = metadata.tables.keys()
+
+    table_objs = [metadata.tables[name] for name in table_names]
+    table_objs = sqlalchemy.sql.util.sort_tables(table_objs)
+
+
     # Drop all tables if requested
     if drop_tables:
         print_start('Dropping tables')
-        metadata.drop_all()
+        for table in reversed(table_objs):
+            table.drop(checkfirst=True)
         print_done()
 
-    metadata.create_all()
+    for table in table_objs:
+        table.create()
     connection = session.connection()
 
     # Okay, run through the tables and actually load the data now
-    for table_obj in metadata.sorted_tables:
+    for table_obj in table_objs:
         table_name = table_obj.name
         insert_stmt = table_obj.insert()
 
@@ -163,7 +212,7 @@ def load(session, directory=None, drop_tables=False, verbose=False):
 
             # Remembering some zillion rows in the session consumes a lot of
             # RAM.  Let's not do that.  Commit every 1000 rows
-            if len(new_rows) > 1000:
+            if len(new_rows) >= 1000:
                 insert_and_commit()
 
         insert_and_commit()
@@ -186,13 +235,16 @@ def load(session, directory=None, drop_tables=False, verbose=False):
 
 
 
-def dump(session, directory=None, verbose=False):
+def dump(session, tables=[], directory=None, verbose=False):
     """Dumps the contents of a database to a set of CSV files.  Probably not
     useful to anyone besides a developer.
 
     `session`
         SQLAlchemy session to use.
 
+    `tables`
+        List of tables to dump.  If omitted, all tables are dumped.
+
     `directory`
         Directory the CSV files should be put in.  Defaults to the `pokedex`
         data directory.
@@ -208,7 +260,16 @@ def dump(session, directory=None, verbose=False):
     if not directory:
         directory = pkg_resources.resource_filename('pokedex', 'data/csv')
 
-    for table_name in sorted(metadata.tables.keys()):
+    if tables:
+        regex = _wildcards_to_regex(tables)
+        table_names = filter(regex.match, metadata.tables.keys())
+    else:
+        table_names = metadata.tables.keys()
+
+    table_names.sort()
+
+
+    for table_name in table_names:
         print_start(table_name)
         table = metadata.tables[table_name]