pokedex/db/load.py

   1 """CSV to database or vice versa."""
   2 import csv
   3 import pkg_resources
   4 import sys
   5
   6 from sqlalchemy.orm.attributes import instrumentation_registry
   7 import sqlalchemy.types
   8
   9 from pokedex.db import metadata
  10 import pokedex.db.tables as tables
  11
  12
  13 def load(session, directory=None, drop_tables=False):
  14     """Load data from CSV files into the given database session.
  15
  16     Tables are created automatically.
  17
  18     `session`
  19         SQLAlchemy session to use.
  20
  21     `directory`
  22         Directory the CSV files reside in.  Defaults to the `pokedex` data
  23         directory.
  24
  25     `drop_tables`
  26         If set to True, existing `pokedex`-related tables will be dropped.
  27     """
  28
  29     if not directory:
  30         directory = pkg_resources.resource_filename('pokedex', 'data/csv')
  31
  32     # Drop all tables if requested
  33     if options.drop_tables:
  34         print 'Dropping tables...'
  35         metadata.drop_all()
  36
  37     metadata.create_all()
  38
  39     # SQLAlchemy is retarded and there is no way for me to get a list of ORM
  40     # classes besides to inspect the module they all happen to live in for
  41     # things that look right.
  42     table_base = tables.TableBase
  43     orm_classes = {}  # table object => table class
  44
  45     for name in dir(tables):
  46         # dir() returns strings!  How /convenient/.
  47         thingy = getattr(tables, name)
  48
  49         if not isinstance(thingy, type):
  50             # Not a class; bail
  51             continue
  52         elif not issubclass(thingy, table_base):
  53             # Not a declarative table; bail
  54             continue
  55         elif thingy == table_base:
  56             # Declarative table base, so not a real table; bail
  57             continue
  58
  59         # thingy is definitely a table class!  Hallelujah.
  60         orm_classes[thingy.__table__] = thingy
  61
  62     # Okay, run through the tables and actually load the data now
  63     for table_obj in metadata.sorted_tables:
  64         table_class = orm_classes[table_obj]
  65         table_name = table_obj.name
  66
  67         # Print the table name but leave the cursor in a fixed column
  68         print table_name + '...', ' ' * (40 - len(table_name)),
  69         sys.stdout.flush()
  70
  71         try:
  72             csvfile = open("%s/%s.csv" % (directory, table_name), 'rb')
  73         except IOError:
  74             # File doesn't exist; don't load anything!
  75             print 'no data!'
  76             continue
  77
  78         reader = csv.reader(csvfile, lineterminator='\n')
  79         column_names = [unicode(column) for column in reader.next()]
  80
  81         # Self-referential tables may contain rows with foreign keys of other
  82         # rows in the same table that do not yet exist.  Pull these out and add
  83         # them to the session last
  84         # ASSUMPTION: Self-referential tables have a single PK called "id"
  85         deferred_rows = []  # ( row referring to id, [foreign ids we need] )
  86         seen_ids = {}       # primary key we've seen => 1
  87
  88         # Fetch foreign key columns that point at this table, if any
  89         self_ref_columns = []
  90         for column in table_obj.c:
  91             if any(_.references(table_obj) for _ in column.foreign_keys):
  92                 self_ref_columns.append(column)
  93
  94         for csvs in reader:
  95             row = table_class()
  96
  97             for column_name, value in zip(column_names, csvs):
  98                 column = table_obj.c[column_name]
  99                 if column.nullable and value == '':
 100                     # Empty string in a nullable column really means NULL
 101                     value = None
 102                 elif isinstance(column.type, sqlalchemy.types.Boolean):
 103                     # Boolean values are stored as string values 0/1, but both
 104                     # of those evaluate as true; SQLA wants True/False
 105                     if value == '0':
 106                         value = False
 107                     else:
 108                         value = True
 109                 else:
 110                     # Otherwise, unflatten from bytes
 111                     value = value.decode('utf-8')
 112
 113                 setattr(row, column_name, value)
 114
 115             # May need to stash this row and add it later if it refers to a
 116             # later row in this table
 117             if self_ref_columns:
 118                 foreign_ids = [getattr(row, _.name) for _ in self_ref_columns]
 119                 foreign_ids = [_ for _ in foreign_ids if _]  # remove NULL ids
 120
 121                 if not foreign_ids:
 122                     # NULL key.  Remember this row and add as usual.
 123                     seen_ids[row.id] = 1
 124
 125                 elif all(_ in seen_ids for _ in foreign_ids):
 126                     # Non-NULL key we've already seen.  Remember it and commit
 127                     # so we know the old row exists when we add the new one
 128                     session.commit()
 129                     seen_ids[row.id] = 1
 130
 131                 else:
 132                     # Non-NULL future id.  Save this and insert it later!
 133                     deferred_rows.append((row, foreign_ids))
 134                     continue
 135
 136             session.add(row)
 137
 138         session.commit()
 139
 140         # Attempt to add any spare rows we've collected
 141         for row, foreign_ids in deferred_rows:
 142             if not all(_ in seen_ids for _ in foreign_ids):
 143                 # Could happen if row A refers to B which refers to C.
 144                 # This is ridiculous and doesn't happen in my data so far
 145                 raise ValueError("Too many levels of self-reference!  "
 146                                  "Row was: " + str(row.__dict__))
 147
 148             session.add(row)
 149             seen_ids[row.id] = 1
 150             session.commit()
 151
 152         print 'loaded'
 153
 154
 155
 156 def dump(session, directory=None):
 157     """Dumps the contents of a database to a set of CSV files.  Probably not
 158     useful to anyone besides a developer.
 159
 160     `session`
 161         SQLAlchemy session to use.
 162
 163     `directory`
 164         Directory the CSV files should be put in.  Defaults to the `pokedex`
 165         data directory.
 166     """
 167
 168     if not directory:
 169         directory = pkg_resources.resource_filename('pokedex', 'data/csv')
 170
 171     for table_name in sorted(metadata.tables.keys()):
 172         print table_name
 173         table = metadata.tables[table_name]
 174
 175         writer = csv.writer(open("%s/%s.csv" % (directory, table_name), 'wb'),
 176                             lineterminator='\n')
 177         columns = [col.name for col in table.columns]
 178         writer.writerow(columns)
 179
 180         primary_key = table.primary_key
 181         for row in session.query(table).order_by(*primary_key).all():
 182             csvs = []
 183             for col in columns:
 184                 # Convert Pythony values to something more universal
 185                 val = getattr(row, col)
 186                 if val == None:
 187                     val = ''
 188                 elif val == True:
 189                     val = '1'
 190                 elif val == False:
 191                     val = '0'
 192                 else:
 193                     val = unicode(val).encode('utf-8')
 194
 195                 csvs.append(val)
 196
 197             writer.writerow(csvs)