Project

General

Profile

1
#!/usr/bin/env python
2
# Maps one datasource to another, using a map spreadsheet if needed
3
# Exit status is the # of errors in the import, up to the maximum exit status
4
# For outputting an XML file to a PostgreSQL database, use the general format of
5
# http://vegbank.org/vegdocs/xml/vegbank_example_ver1.0.2.xml
6

    
7
import copy
8
import csv
9
import itertools
10
import os.path
11
import sys
12
import xml.dom.minidom as minidom
13

    
14
sys.path.append(os.path.dirname(__file__)+"/../lib")
15

    
16
import csvs
17
import db_xml
18
import exc
19
import iters
20
import maps
21
import opts
22
import parallelproc
23
import Parser
24
import profiling
25
import sql
26
import sql_gen
27
import sql_io
28
import streams
29
import strings
30
import term
31
import util
32
import xpath
33
import xml_dom
34
import xml_func
35
import xml_parse
36

    
37
def get_with_prefix(map_, prefixes, key):
38
    '''Gets all entries for the given key with any of the given prefixes
39
    @return tuple(found_key, found_value)
40
    '''
41
    values = []
42
    for key_ in strings.with_prefixes(['']+prefixes, key): # also with no prefix
43
        try: value = map_[key_]
44
        except KeyError, e: continue # keep going
45
        values.append((key_, value))
46
    
47
    if values != []: return values
48
    else: raise e # re-raise last KeyError
49

    
50
def metadata_value(name): return None # this feature has been removed
51

    
52
def cleanup(val):
53
    if val == None: return val
54
    return util.none_if(strings.cleanup(strings.ustr(val)), u'', u'\\N')
55

    
56
def main_():
57
    env_names = []
58
    def usage_err():
59
        raise SystemExit('Usage: '+opts.env_usage(env_names, True)+' '
60
            +sys.argv[0]+' [map_path...] [<input] [>output]\n'
61
            'Note: Row #s start with 1')
62
    
63
    ## Get config from env vars
64
    
65
    # Modes
66
    test = opts.env_flag('test', False, env_names)
67
    commit = opts.env_flag('commit', False, env_names) and not test
68
        # never commit in test mode
69
    redo = opts.env_flag('redo', test, env_names) and not commit
70
        # never redo in commit mode (run `make schemas/reinstall` instead)
71
    
72
    # Ranges
73
    start = util.cast(int, opts.get_env_var('start', 1, env_names)) # 1-based
74
    # Make start interally 0-based.
75
    # It's 1-based to the user to match up with the staging table row #s.
76
    start -= 1
77
    if test: n_default = 1
78
    else: n_default = None
79
    n = util.cast(int, util.none_if(opts.get_env_var('n', n_default, env_names),
80
        u''))
81
    end = n
82
    if end != None: end += start
83
    
84
    # Debugging
85
    verbosity = util.cast(float, opts.get_env_var('verbosity', None, env_names))
86
    opts.get_env_var('profile_to', None, env_names) # add to env_names
87
    
88
    # DB
89
    def get_db_config(prefix):
90
        return opts.get_env_vars(sql.db_config_names, prefix, env_names)
91
    in_db_config = get_db_config('in')
92
    out_db_config = get_db_config('out')
93
    in_is_db = 'engine' in in_db_config
94
    out_is_db = 'engine' in out_db_config
95
    in_schema = opts.get_env_var('in_schema', None, env_names)
96
    in_table = opts.get_env_var('in_table', None, env_names)
97
    if in_schema != None:
98
        for config in [in_db_config, out_db_config]:
99
            config['schemas'] += ','+in_schema
100
    
101
    # Optimization
102
    cache_sql = opts.env_flag('cache_sql', True, env_names)
103
    by_col = in_db_config == out_db_config and opts.env_flag('by_col', False,
104
        env_names) # by-column optimization only applies if mapping to same DB
105
    if test: cpus_default = 0
106
    else: cpus_default = 0 # or None to use parallel processing by default
107
    cpus = util.cast(int, util.none_if(opts.get_env_var('cpus', cpus_default,
108
        env_names), u''))
109
    
110
    # Set default verbosity. Must happen after by_col is set.
111
    if verbosity == None: 
112
        if test: verbosity = 0.5 # automated tests should not be verbose
113
        elif by_col: verbosity = 3 # show all queries to assist profiling
114
        else: verbosity = 1.1 # just show row progress
115
    
116
    ##
117
    
118
    # Logging
119
    verbose_errors = test and verbosity > 0
120
    debug = verbosity >= 1.5
121
    def log(msg, level=1):
122
        '''Higher level -> more verbose'''
123
        if level <= verbosity:
124
            if verbosity <= 2:
125
                if level == 1.5: msg = '# '+msg # msg is Redmine list item
126
                elif msg.startswith('DB query:'): # remove extra debug info
127
                    first_line, nl, msg = msg.partition('\n')
128
            elif level > 1: msg = '['+str(level)+'] '+msg # include level in msg
129
            
130
            sys.stderr.write(strings.to_raw_str(msg.rstrip('\n')+'\n'))
131
    if debug: log_debug = lambda msg, level=2: log(msg, level)
132
    else: log_debug = sql.log_debug_none
133
    
134
    # Parse args
135
    map_paths = sys.argv[1:]
136
    if map_paths == []:
137
        if in_is_db or not out_is_db: usage_err()
138
        else: map_paths = [None]
139
    
140
    def connect_db(db_config):
141
        log('Connecting to '+sql.db_config_str(db_config))
142
        return sql.connect(db_config, caching=cache_sql, autocommit=commit,
143
            debug_temp=verbosity > 3 and commit, log_debug=log_debug)
144
    
145
    if end != None: end_str = str(end-1) # end is one past the last #
146
    else: end_str = 'end'
147
    log('Processing input rows '+str(start)+'-'+end_str)
148
    
149
    ex_tracker = exc.ExPercentTracker(iter_text='row')
150
    profiler = profiling.ItersProfiler(start_now=True, iter_text='row')
151
    
152
    # Parallel processing
153
    pool = parallelproc.MultiProducerPool(cpus)
154
    log('Using '+str(pool.process_ct)+' parallel CPUs')
155
    
156
    doc = xml_dom.create_doc()
157
    root = doc.documentElement
158
    out_is_xml_ref = [False]
159
    in_label_ref = [None]
160
    def update_in_label():
161
        if in_label_ref[0] != None:
162
            xpath.get(root, '/_ignore/inLabel="'+in_label_ref[0]+'"', True)
163
    def prep_root():
164
        root.clear()
165
        update_in_label()
166
    prep_root()
167
    
168
    # Define before the out_is_db section because it's used by by_col
169
    row_ins_ct_ref = [0]
170
    
171
    if out_is_db:
172
        out_db = connect_db(out_db_config)
173
        rel_funcs = set(sql.tables(out_db, schema_like='%',
174
            table_like=r'\__%'))
175
    
176
    def process_input(root, row_ready, map_path):
177
        '''Inputs datasource to XML tree, mapping if needed'''
178
        # Load map header
179
        in_is_xpaths = True
180
        out_is_xpaths = True
181
        out_label = None
182
        if map_path != None:
183
            metadata = []
184
            mappings = []
185
            stream = open(map_path, 'rb')
186
            reader = csv.reader(stream)
187
            in_label, out_label = reader.next()[:2]
188
            
189
            def split_col_name(name):
190
                label, sep, root = name.partition(':')
191
                label, sep2, prefixes_str = label.partition('[')
192
                prefixes_str = strings.remove_suffix(']', prefixes_str)
193
                prefixes = strings.split(',', prefixes_str)
194
                return label, sep != '', root, prefixes
195
                    # extract datasrc from "datasrc[data_format]"
196
            
197
            in_label, in_root, prefixes = maps.col_info(in_label)
198
            in_is_xpaths = in_root != None
199
            in_label_ref[0] = in_label
200
            update_in_label()
201
            out_label, out_root = maps.col_info(out_label)[:2]
202
            out_is_xpaths = out_root != None
203
            if out_is_xpaths: has_types = out_root.find('/*s/') >= 0
204
                # outer elements are types
205
            
206
            for row in reader:
207
                in_, out = row[:2]
208
                if out != '': mappings.append([in_, out_root+out])
209
            
210
            stream.close()
211
            
212
            root.ownerDocument.documentElement.tagName = out_label
213
        in_is_xml = in_is_xpaths and not in_is_db
214
        out_is_xml_ref[0] = out_is_xpaths and not out_is_db
215
        
216
        def process_rows(process_row, rows, rows_start=0):
217
            '''Processes input rows      
218
            @param process_row(in_row, i)
219
            @rows_start The (0-based) row # of the first row in rows. Set this
220
                only if the pre-start rows have already been skipped.
221
            '''
222
            rows = iter(rows)
223
            
224
            if end != None: row_nums = xrange(rows_start, end)
225
            else: row_nums = itertools.count(rows_start)
226
            i = -1
227
            for i in row_nums:
228
                try: row = rows.next()
229
                except StopIteration:
230
                    i -= 1 # last row # didn't count
231
                    break # no more rows
232
                if i < start: continue # not at start row yet
233
                
234
                # Row # is interally 0-based, but 1-based to the user
235
                log('Processing input row #'+str(i+1), level=1.1)
236
                process_row(row, i)
237
                row_ready(i, row)
238
            row_ct = i-start+1
239
            return row_ct
240
        
241
        def map_rows(get_value, rows, **kw_args):
242
            '''Maps input rows
243
            @param get_value(in_, row):str
244
            '''
245
            # Prevent collisions if multiple inputs mapping to same output
246
            outputs_idxs = dict()
247
            for i, mapping in enumerate(mappings):
248
                in_, out = mapping
249
                default = util.NamedTuple(count=1, first=i)
250
                idxs = outputs_idxs.setdefault(out, default)
251
                if idxs is not default: # key existed, so there was a collision
252
                    if idxs.count == 1: # first key does not yet have /_alt/#
253
                        mappings[idxs.first][1] += '/_alt/0'
254
                    mappings[i][1] += '/_alt/'+str(idxs.count)
255
                    idxs.count += 1
256
            
257
            id_node = None
258
            if out_is_db:
259
                for i, mapping in enumerate(mappings):
260
                    in_, out = mapping
261
                    # All put_obj()s should return the same id_node
262
                    nodes, id_node = xpath.put_obj(root, out, '-1', has_types,
263
                        '$'+str(in_)) # value is placeholder that documents name
264
                    mappings[i] = [in_, nodes]
265
                assert id_node != None
266
                
267
                if debug: # only calc if debug
268
                    log_debug('Put template:\n'+str(root))
269
            
270
            def process_row(row, i):
271
                row_id = str(i)
272
                if id_node != None: xml_dom.set_value(id_node, row_id)
273
                for in_, out in mappings:
274
                    log_debug('Getting '+str(in_))
275
                    value = metadata_value(in_)
276
                    if value == None: value = cleanup(get_value(in_, row))
277
                    log_debug('Putting '+repr(value)+' to '+str(out))
278
                    if out_is_db: # out is list of XML nodes
279
                        for node in out: xml_dom.set_value(node, value)
280
                    elif value != None: # out is XPath
281
                        xpath.put_obj(root, out, row_id, has_types, value)
282
            return process_rows(process_row, rows, **kw_args)
283
        
284
        def map_table(col_names, rows, **kw_args):
285
            col_names_ct = len(col_names)
286
            col_idxs = util.list_flip(col_names)
287
            
288
            # Resolve prefixes
289
            mappings_orig = mappings[:] # save a copy
290
            mappings[:] = [] # empty existing elements
291
            for in_, out in mappings_orig:
292
                if metadata_value(in_) == None:
293
                    try: cols = get_with_prefix(col_idxs, prefixes, in_)
294
                    except KeyError: pass
295
                    else: mappings[len(mappings):] = [[db_xml.ColRef(*col), out]
296
                        for col in cols] # can't use += because that uses =
297
            
298
            def get_value(in_, row): return row.list[in_.idx]
299
            def wrap_row(row):
300
                return util.ListDict(util.list_as_length(row, col_names_ct),
301
                    col_names, col_idxs) # handle CSV rows of different lengths
302
            
303
            return map_rows(get_value, util.WrapIter(wrap_row, rows), **kw_args)
304
        
305
        if in_is_db:
306
            def on_error(e): ex_tracker.track(e)
307
            
308
            if by_col: in_db = out_db
309
            else: in_db = connect_db(in_db_config)
310
            
311
            # Get table and schema name
312
            schema = in_schema # modified, so can't have same name as outer var
313
            table = in_table # modified, so can't have same name as outer var
314
            if table == None:
315
                assert in_is_xpaths
316
                schema, sep, table = in_root.partition('.')
317
                if sep == '': # only the table name was specified
318
                    table = schema
319
                    schema = None
320
            table = sql_gen.Table(table, schema)
321
            
322
            # Fetch rows
323
            if by_col: limit = 0 # only fetch column names
324
            else: limit = n
325
            cur = sql.select(in_db, table, limit=limit, start=start,
326
                cacheable=False)
327
            col_names = list(sql.col_names(cur))
328
            
329
            if by_col:
330
                map_table(col_names, []) # just create the template
331
                
332
                if n == None: # doing full import
333
                    log('Clearing errors table')
334
                    sql.truncate(in_db, sql_io.errors_table(in_db, table))
335
                
336
                # Strip XML functions not in the DB
337
                special_funcs = db_xml.put_table_special_funcs | rel_funcs
338
                xml_func.process(root, rel_funcs=special_funcs)
339
                if debug: log_debug('Putting stripped:\n'+str(root))
340
                    # only calc if debug
341
                
342
                # Import rows
343
                in_row_ct_ref = [0]
344
                db_xml.put_table(in_db, root.firstChild, table, in_row_ct_ref,
345
                    row_ins_ct_ref, n, start, on_error)
346
                row_ct = in_row_ct_ref[0]
347
            else:
348
                # Use normal by-row method
349
                row_ct = map_table(col_names, sql.rows(cur), rows_start=start)
350
                    # rows_start: pre-start rows have been skipped
351
                
352
                in_db.db.close()
353
        elif in_is_xml:
354
            stdin = streams.LineCountStream(sys.stdin)
355
            def on_error(e):
356
                exc.add_msg(e, term.emph('input line #:')+' '
357
                    +str(stdin.line_num))
358
                ex_tracker.track(e)
359
            
360
            def get_rows(doc2rows):
361
                return iters.flatten(itertools.imap(doc2rows,
362
                    xml_parse.docs_iter(stdin, on_error)))
363
            
364
            if map_path == None:
365
                def doc2rows(in_xml_root):
366
                    iter_ = xml_dom.NodeElemIter(in_xml_root)
367
                    util.skip(iter_, xml_dom.is_text) # skip metadata
368
                    return iter_
369
                
370
                row_ct = process_rows(lambda row, i: root.appendChild(row),
371
                    get_rows(doc2rows))
372
            else:
373
                def doc2rows(in_xml_root):
374
                    rows = xpath.get(in_xml_root, in_root, limit=end)
375
                    if rows == []: raise SystemExit('Map error: Root "'
376
                        +in_root+'" not found in input')
377
                    return rows
378
                
379
                def get_value(in_, row):
380
                    in_ = './{'+(','.join(strings.with_prefixes(
381
                        ['']+prefixes, in_)))+'}' # also with no prefix
382
                    nodes = xpath.get(row, in_, allow_rooted=False)
383
                    if nodes != []: return xml_dom.value(nodes[0])
384
                    else: return None
385
                
386
                row_ct = map_rows(get_value, get_rows(doc2rows))
387
        else: # input is CSV
388
            map_ = dict(mappings)
389
            reader, col_names = csvs.reader_and_header(sys.stdin)
390
            row_ct = map_table(col_names, reader)
391
        
392
        return row_ct
393
    
394
    def process_inputs(root, row_ready):
395
        row_ct = 0
396
        for map_path in map_paths:
397
            row_ct += process_input(root, row_ready, map_path)
398
        return row_ct
399
    
400
    pool.share_vars(locals())
401
    if out_is_db:
402
        try:
403
            if redo: sql.empty_db(out_db)
404
            pool.share_vars(locals())
405
            
406
            def row_ready(row_num, input_row):
407
                row_str_ = [None]
408
                def row_str():
409
                    if row_str_[0] == None:
410
                        # Row # is interally 0-based, but 1-based to the user
411
                        row_str_[0] = (term.emph('row #:')+' '+str(row_num+1)
412
                            +'\n'+term.emph('input row:')+'\n'+str(input_row))
413
                        if verbose_errors: row_str_[0] += ('\n'
414
                            +term.emph('output row:')+'\n'+str(root))
415
                    return row_str_[0]
416
                
417
                if debug: log_debug(row_str()) # only calc if debug
418
                
419
                def on_error(e):
420
                    exc.add_msg(e, row_str())
421
                    ex_tracker.track(e, row_num, detail=verbose_errors)
422
                pool.share_vars(locals())
423
                
424
                row_root = root.cloneNode(True) # deep copy so don't modify root
425
                xml_func.process(row_root, on_error, rel_funcs, out_db)
426
                if not xml_dom.is_empty(row_root):
427
                    assert xml_dom.has_one_child(row_root)
428
                    try:
429
                        sql.with_savepoint(out_db,
430
                            lambda: db_xml.put(out_db, row_root.firstChild,
431
                                row_ins_ct_ref, on_error))
432
                    except sql.DatabaseErrors, e: on_error(e)
433
            
434
            row_ct = process_inputs(root, row_ready)
435
            sys.stdout.write('Inserted '+str(row_ins_ct_ref[0])+
436
                ' new rows into database\n')
437
            
438
            # Consume asynchronous tasks
439
            pool.main_loop()
440
        finally: out_db.close()
441
    else:
442
        def on_error(e): ex_tracker.track(e)
443
        def row_ready(row_num, input_row): pass
444
        row_ct = process_inputs(root, row_ready)
445
        xml_func.process(root, on_error)
446
        if out_is_xml_ref[0]:
447
            doc.writexml(sys.stdout, **xml_dom.prettyxml_config)
448
        else: # output is CSV
449
            raise NotImplementedError('CSV output not supported yet')
450
    
451
    # Consume any asynchronous tasks not already consumed above
452
    pool.main_loop()
453
    
454
    profiler.stop(row_ct)
455
    if not by_col: ex_tracker.add_iters(row_ct) # only if errors are done by row
456
    log('Processed '+str(row_ct)+' input rows')
457
    log(profiler.msg())
458
    log(ex_tracker.msg())
459
    ex_tracker.exit()
460

    
461
def main():
462
    try: main_()
463
    except Parser.SyntaxError, e: raise SystemExit(str(e))
464

    
465
if __name__ == '__main__':
466
    profile_to = opts.get_env_var('profile_to', None)
467
    if profile_to != None:
468
        import cProfile
469
        sys.stderr.write('Profiling to '+profile_to+'\n')
470
        cProfile.run(main.func_code, profile_to)
471
    else: main()
(26-26/52)