Project

General

Profile

1 996 aaronmk
# XML "function" nodes that transform their contents
2 86 aaronmk
3 111 aaronmk
import datetime
4 968 aaronmk
import re
5 1219 aaronmk
import sre_constants
6 2017 aaronmk
import warnings
7 111 aaronmk
8 1607 aaronmk
import angles
9 818 aaronmk
import dates
10 300 aaronmk
import exc
11 1580 aaronmk
import format
12 917 aaronmk
import maps
13 3077 aaronmk
import sql_io
14 1234 aaronmk
import strings
15 827 aaronmk
import term
16 1468 aaronmk
import units
17 1047 aaronmk
import util
18 86 aaronmk
import xml_dom
19 1321 aaronmk
import xpath
20 86 aaronmk
21 995 aaronmk
##### Exceptions
22
23 1612 aaronmk
class SyntaxError(exc.ExceptionWithCause):
24 797 aaronmk
    def __init__(self, cause):
25 1611 aaronmk
        exc.ExceptionWithCause.__init__(self, 'Invalid XML function syntax',
26
            cause)
27 278 aaronmk
28 1613 aaronmk
class FormatException(exc.ExceptionWithCause):
29
    def __init__(self, cause):
30
        exc.ExceptionWithCause.__init__(self, 'Invalid input value', cause)
31 843 aaronmk
32 1992 aaronmk
##### Helper functions
33 995 aaronmk
34 1992 aaronmk
def map_items(func, items):
35
    return [(name, func(value)) for name, value in items]
36
37
def cast(type_, val):
38
    '''Throws FormatException if can't cast'''
39
    try: return type_(val)
40
    except ValueError, e: raise FormatException(e)
41
42
def conv_items(type_, items):
43
    return map_items(lambda val: cast(type_, val),
44
        xml_dom.TextEntryOnlyIter(items))
45
46
def pop_value(items, name='value'):
47
    '''@param name Name of value param, or None to accept any name'''
48
    try: last = items.pop() # last entry contains value
49
    except IndexError: return None # input is empty and no actions
50
    if name != None and last[0] != name: return None # input is empty
51
    return last[1]
52
53 995 aaronmk
funcs = {}
54
55 2557 aaronmk
structural_funcs = set()
56
57 1992 aaronmk
##### Public functions
58
59 2112 aaronmk
def is_func_name(name):
60
    return name.startswith('_') and name != '_' # '_' is default root node name
61
62
def is_func(node): return is_func_name(node.tagName)
63
64
def is_xml_func_name(name): return is_func_name(name) and name in funcs
65
66
def is_xml_func(node): return is_xml_func_name(node.tagName)
67
68 2602 aaronmk
def process(node, on_error=exc.raise_, rel_funcs=None, db=None):
69 2597 aaronmk
    '''Evaluates the XML functions in an XML tree.
70 2602 aaronmk
    @param rel_funcs None|set(str...) Relational functions
71 2597 aaronmk
        * container can be any iterable type
72 2602 aaronmk
        * If != None: Non-relational functions are removed, or relational
73
          functions are treated specially, depending on the db param (below).
74
    @param db
75
        * If None: Non-relational functions other than structural functions are
76
          replaced with their last parameter (usually the value), not evaluated.
77
          This is used in column-based mode to remove XML-only functions.
78
        * If != None: Relational functions are evaluated directly. This is used
79
          in row-based mode to combine relational and XML functions.
80 2597 aaronmk
    '''
81 2602 aaronmk
    has_rel_funcs = rel_funcs != None
82
    assert db == None or has_rel_funcs # rel_funcs required if db set
83 2597 aaronmk
84
    for child in xml_dom.NodeElemIter(node):
85 2602 aaronmk
        process(child, on_error, rel_funcs, db)
86
87 995 aaronmk
    name = node.tagName
88 2602 aaronmk
    if not is_func_name(name): return # not any kind of function
89
90
    # Change rel_funcs *after* processing child nodes, which needs orig value
91
    if not has_rel_funcs: rel_funcs = set()
92
    rel_funcs = set(rel_funcs)
93
94
    row_mode = has_rel_funcs and db != None
95
    column_mode = has_rel_funcs and db == None
96 3028 aaronmk
    items = list(xml_dom.NodeTextEntryIter(node))
97 2602 aaronmk
98 3029 aaronmk
    # Parse function
99
    if len(items) == 1 and items[0][0].isdigit(): # has single numeric param
100
        # pass-through optimization for aggregating functions with one arg
101
        value = items[0][1] # pass through first arg
102
    elif row_mode and name in rel_funcs: # row-based mode: evaluate using DB
103 3077 aaronmk
        value = sql_io.put(db, name, dict(items))
104 2602 aaronmk
    elif column_mode and not name in structural_funcs: # column-based mode
105
        if name in rel_funcs: return # preserve relational functions
106
        # otherwise XML-only, so just replace with last param
107 3028 aaronmk
        value = pop_value(items, None)
108 2602 aaronmk
    else: # local XML function
109
        try: value = funcs[name](items, node)
110 1613 aaronmk
        except Exception, e: # also catch non-wrapped exceptions (XML func bugs)
111 1371 aaronmk
            # Save in case another exception raised, overwriting sys.exc_info()
112
            exc.add_traceback(e)
113 1562 aaronmk
            str_ = strings.ustr(node)
114 995 aaronmk
            exc.add_msg(e, 'function:\n'+str_)
115 1810 aaronmk
            xml_dom.replace(node, xml_dom.mk_comment(node.ownerDocument,
116
                '\n'+term.emph_multiline(str_)))
117
118 995 aaronmk
            on_error(e)
119 2602 aaronmk
            return # in case on_error() returns
120
    xml_dom.replace_with_text(node, value)
121 995 aaronmk
122 1469 aaronmk
##### XML functions
123 995 aaronmk
124
# Function names must start with _ to avoid collisions with real tags
125
# Functions take arguments (items)
126
127 2557 aaronmk
#### Structural
128 1469 aaronmk
129 2017 aaronmk
def _ignore(items, node):
130 994 aaronmk
    '''Used to "comment out" an XML subtree'''
131
    return None
132 995 aaronmk
funcs['_ignore'] = _ignore
133 2557 aaronmk
structural_funcs.add('_ignore')
134 994 aaronmk
135 2017 aaronmk
def _ref(items, node):
136
    '''Used to retrieve a value from another XML node
137
    @param items
138
        addr=<path> XPath to value, relative to the XML func's parent node
139
    '''
140
    items = dict(items)
141
    try: addr = items['addr']
142
    except KeyError, e: raise SyntaxError(e)
143
144
    value = xpath.get_value(node.parentNode, addr)
145
    if value == None:
146
        warnings.warn(UserWarning('_ref: XPath reference target missing: '
147
            +str(addr)))
148
    return value
149
funcs['_ref'] = _ref
150 2557 aaronmk
structural_funcs.add('_ref')
151 2017 aaronmk
152 1469 aaronmk
#### Conditionals
153
154 2016 aaronmk
def _eq(items, node):
155 1234 aaronmk
    items = dict(items)
156
    try:
157
        left = items['left']
158
        right = items['right']
159
    except KeyError: return '' # a value was None
160
    return util.bool2str(left == right)
161
funcs['_eq'] = _eq
162
163 2016 aaronmk
def _if(items, node):
164 1234 aaronmk
    items = dict(items)
165
    try:
166
        cond = items['cond']
167
        then = items['then']
168 1612 aaronmk
    except KeyError, e: raise SyntaxError(e)
169 1234 aaronmk
    else_ = items.get('else', None)
170 1562 aaronmk
    cond = bool(cast(strings.ustr, cond))
171 1234 aaronmk
    if cond: return then
172
    else: return else_
173
funcs['_if'] = _if
174
175 1469 aaronmk
#### Combining values
176
177 2016 aaronmk
def _alt(items, node):
178 113 aaronmk
    items = list(items)
179
    items.sort()
180 1186 aaronmk
    try: return items[0][1] # value of lowest-numbered item
181 1609 aaronmk
    except IndexError: return None # input got removed by e.g. FormatException
182 995 aaronmk
funcs['_alt'] = _alt
183 113 aaronmk
184 2016 aaronmk
def _merge(items, node):
185 1234 aaronmk
    items = list(conv_items(strings.ustr, items))
186 1562 aaronmk
        # get *once* from iter, check types
187 917 aaronmk
    items.sort()
188
    return maps.merge_values(*[v for k, v in items])
189 995 aaronmk
funcs['_merge'] = _merge
190 917 aaronmk
191 2016 aaronmk
def _label(items, node):
192 1412 aaronmk
    items = dict(conv_items(strings.ustr, items))
193 1562 aaronmk
        # get *once* from iter, check types
194 2014 aaronmk
    value = items.get('value', None)
195
    if value == None: return None # input is empty
196
    try: label = items['label']
197 1612 aaronmk
    except KeyError, e: raise SyntaxError(e)
198 917 aaronmk
    return label+': '+value
199 995 aaronmk
funcs['_label'] = _label
200 917 aaronmk
201 1469 aaronmk
#### Transforming values
202
203 2016 aaronmk
def _collapse(items, node):
204 2012 aaronmk
    '''Collapses a subtree if the "value" element in it is NULL'''
205
    items = dict(items)
206
    try: require = cast(strings.ustr, items['require'])
207
    except KeyError, e: raise SyntaxError(e)
208
    value = items.get('value', None)
209
210 2558 aaronmk
    if xpath.get_value(value, require, allow_rooted=False) == None: return None
211 2012 aaronmk
    else: return value
212
funcs['_collapse'] = _collapse
213
214 1478 aaronmk
types_by_name = {None: strings.ustr, 'str': strings.ustr, 'float': float}
215 1477 aaronmk
216 2016 aaronmk
def _nullIf(items, node):
217 1562 aaronmk
    items = dict(conv_items(strings.ustr, items))
218 1477 aaronmk
    try: null = items['null']
219 1612 aaronmk
    except KeyError, e: raise SyntaxError(e)
220 1477 aaronmk
    value = items.get('value', None)
221 1219 aaronmk
    type_str = items.get('type', None)
222 1477 aaronmk
223
    try: type_ = types_by_name[type_str]
224 1612 aaronmk
    except KeyError, e: raise SyntaxError(e)
225 1477 aaronmk
    null = type_(null)
226
227
    try: return util.none_if(value, null)
228
    except ValueError: return value # value not convertible, so can't equal null
229 1047 aaronmk
funcs['_nullIf'] = _nullIf
230
231 1602 aaronmk
def repl(repls, value):
232 1537 aaronmk
    '''Raises error if value not in map and no special '*' entry
233 1602 aaronmk
    @param repls dict repl:with
234
        repl "*" means all other input values
235
        with "*" means keep input value the same
236
        with "" means ignore input value
237 1537 aaronmk
    '''
238 1602 aaronmk
    try: new_value = repls[value]
239 1304 aaronmk
    except KeyError, e:
240 1537 aaronmk
        # Save traceback right away in case another exception raised
241 2984 aaronmk
        fe = FormatException(e)
242 1602 aaronmk
        try: new_value = repls['*']
243 1609 aaronmk
        except KeyError: raise fe
244 1537 aaronmk
    if new_value == '*': new_value = value # '*' means keep input value the same
245 1607 aaronmk
    return new_value
246 1602 aaronmk
247 2016 aaronmk
def _map(items, node):
248 1602 aaronmk
    '''See repl()
249
    @param items
250
        <last_entry> Value
251
        <other_entries> name=value Mappings. Special values: See repl() repls.
252
    '''
253
    items = conv_items(strings.ustr, items) # get *once* from iter, check types
254
    value = pop_value(items)
255
    if value == None: return None # input is empty
256 1607 aaronmk
    return util.none_if(repl(dict(items), value), u'') # empty value means None
257 1219 aaronmk
funcs['_map'] = _map
258
259 2016 aaronmk
def _replace(items, node):
260 1562 aaronmk
    items = conv_items(strings.ustr, items) # get *once* from iter, check types
261 1581 aaronmk
    value = pop_value(items)
262
    if value == None: return None # input is empty
263 1219 aaronmk
    try:
264
        for repl, with_ in items:
265
            if re.match(r'^\w+$', repl):
266
                repl = r'(?<![^\W_])'+repl+r'(?![^\W_])' # match whole word
267
            value = re.sub(repl, with_, value)
268 1612 aaronmk
    except sre_constants.error, e: raise SyntaxError(e)
269 1624 aaronmk
    return util.none_if(value.strip(), u'') # empty strings always mean None
270 1219 aaronmk
funcs['_replace'] = _replace
271
272 1469 aaronmk
#### Quantities
273
274 2016 aaronmk
def _units(items, node):
275 1562 aaronmk
    items = conv_items(strings.ustr, items) # get *once* from iter, check types
276 1581 aaronmk
    value = pop_value(items)
277
    if value == None: return None # input is empty
278 1471 aaronmk
279 1581 aaronmk
    quantity = units.str2quantity(value)
280 1471 aaronmk
    try:
281
        for action, units_ in items:
282
            units_ = util.none_if(units_, u'')
283
            if action == 'default': units.set_default_units(quantity, units_)
284 1567 aaronmk
            elif action == 'to':
285
                try: quantity = units.convert(quantity, units_)
286 1609 aaronmk
                except ValueError, e: raise FormatException(e)
287 1612 aaronmk
            else: raise SyntaxError(ValueError('Invalid action: '+action))
288 1609 aaronmk
    except units.MissingUnitsException, e: raise FormatException(e)
289 1471 aaronmk
    return units.quantity2str(quantity)
290 1225 aaronmk
funcs['_units'] = _units
291
292 1399 aaronmk
def parse_range(str_, range_sep='-'):
293
    default = (str_, None)
294
    start, sep, end = str_.partition(range_sep)
295
    if sep == '': return default # not a range
296 1427 aaronmk
    if start == '' and range_sep == '-': return default # negative number
297 1399 aaronmk
    return tuple(d.strip() for d in (start, end))
298
299 2016 aaronmk
def _rangeStart(items, node):
300 1562 aaronmk
    items = dict(conv_items(strings.ustr, items))
301 1399 aaronmk
    try: value = items['value']
302 1406 aaronmk
    except KeyError: return None # input is empty
303 1399 aaronmk
    return parse_range(value)[0]
304
funcs['_rangeStart'] = _rangeStart
305
306 2016 aaronmk
def _rangeEnd(items, node):
307 1562 aaronmk
    items = dict(conv_items(strings.ustr, items))
308 1399 aaronmk
    try: value = items['value']
309 1406 aaronmk
    except KeyError: return None # input is empty
310 1399 aaronmk
    return parse_range(value)[1]
311
funcs['_rangeEnd'] = _rangeEnd
312
313 2016 aaronmk
def _range(items, node):
314 1472 aaronmk
    items = dict(conv_items(float, items))
315
    from_ = items.get('from', None)
316
    to = items.get('to', None)
317
    if from_ == None or to == None: return None
318
    return str(to - from_)
319
funcs['_range'] = _range
320
321 2016 aaronmk
def _avg(items, node):
322 86 aaronmk
    count = 0
323
    sum_ = 0.
324 278 aaronmk
    for name, value in conv_items(float, items):
325 86 aaronmk
        count += 1
326
        sum_ += value
327 1472 aaronmk
    if count == 0: return None # input is empty
328
    else: return str(sum_/count)
329 995 aaronmk
funcs['_avg'] = _avg
330 86 aaronmk
331 968 aaronmk
class CvException(Exception):
332
    def __init__(self):
333
        Exception.__init__(self, 'CV (coefficient of variation) values are only'
334
            ' allowed for ratio scale data '
335
            '(see <http://en.wikipedia.org/wiki/Coefficient_of_variation>)')
336
337 2016 aaronmk
def _noCV(items, node):
338 3046 aaronmk
    try: name, value = items.pop() # last entry contains value
339
    except IndexError: return None # input is empty
340 1609 aaronmk
    if re.match('^(?i)CV *\d+$', value): raise FormatException(CvException())
341 968 aaronmk
    return value
342 995 aaronmk
funcs['_noCV'] = _noCV
343 968 aaronmk
344 1469 aaronmk
#### Dates
345
346 2016 aaronmk
def _date(items, node):
347 1562 aaronmk
    items = dict(conv_items(strings.ustr, items))
348
        # get *once* from iter, check types
349 1514 aaronmk
    try: str_ = items['date']
350 786 aaronmk
    except KeyError:
351 1515 aaronmk
        # Year is required
352
        try: items['year']
353 1309 aaronmk
        except KeyError, e:
354
            if items == {}: return None # entire date is empty
355 1609 aaronmk
            else: raise FormatException(e)
356 1515 aaronmk
357
        # Convert month name to number
358
        try: month = items['month']
359
        except KeyError: pass
360
        else:
361
            if not month.isdigit(): # month is name
362 1582 aaronmk
                try: items['month'] = str(dates.strtotime(month).month)
363 1609 aaronmk
                except ValueError, e: raise FormatException(e)
364 1515 aaronmk
365 1580 aaronmk
        items = dict(conv_items(format.str2int, items.iteritems()))
366 786 aaronmk
        items.setdefault('month', 1)
367
        items.setdefault('day', 1)
368 1535 aaronmk
369
        for try_num in xrange(2):
370
            try:
371
                date = datetime.date(**items)
372
                break
373
            except ValueError, e:
374 1609 aaronmk
                if try_num > 0: raise FormatException(e)
375 1536 aaronmk
                    # exception still raised after retry
376 1562 aaronmk
                msg = strings.ustr(e)
377 1535 aaronmk
                if msg == 'month must be in 1..12': # try swapping month and day
378
                    items['month'], items['day'] = items['day'], items['month']
379 1609 aaronmk
                else: raise FormatException(e)
380 786 aaronmk
    else:
381 324 aaronmk
        try: year = float(str_)
382
        except ValueError:
383 1264 aaronmk
            try: date = dates.strtotime(str_)
384 324 aaronmk
            except ImportError: return str_
385 1609 aaronmk
            except ValueError, e: raise FormatException(e)
386 324 aaronmk
        else: date = (datetime.date(int(year), 1, 1) +
387
            datetime.timedelta(round((year % 1.)*365)))
388 818 aaronmk
    try: return dates.strftime('%Y-%m-%d', date)
389 843 aaronmk
    except ValueError, e: raise FormatException(e)
390 995 aaronmk
funcs['_date'] = _date
391 86 aaronmk
392 2016 aaronmk
def _dateRangeStart(items, node):
393 1562 aaronmk
    items = dict(conv_items(strings.ustr, items))
394 1366 aaronmk
    try: value = items['value']
395 1406 aaronmk
    except KeyError: return None # input is empty
396 1366 aaronmk
    return dates.parse_date_range(value)[0]
397
funcs['_dateRangeStart'] = _dateRangeStart
398 1311 aaronmk
399 2016 aaronmk
def _dateRangeEnd(items, node):
400 1562 aaronmk
    items = dict(conv_items(strings.ustr, items))
401 1366 aaronmk
    try: value = items['value']
402 1406 aaronmk
    except KeyError: return None # input is empty
403 1366 aaronmk
    return dates.parse_date_range(value)[1]
404
funcs['_dateRangeEnd'] = _dateRangeEnd
405 1311 aaronmk
406 1469 aaronmk
#### Names
407
408 328 aaronmk
_name_parts_slices_items = [
409
    ('first', slice(None, 1)),
410
    ('middle', slice(1, -1)),
411
    ('last', slice(-1, None)),
412
]
413
name_parts_slices = dict(_name_parts_slices_items)
414
name_parts = [name for name, slice_ in _name_parts_slices_items]
415
416 2016 aaronmk
def _name(items, node):
417 89 aaronmk
    items = dict(items)
418 102 aaronmk
    parts = []
419 328 aaronmk
    for part in name_parts:
420
        if part in items: parts.append(items[part])
421 102 aaronmk
    return ' '.join(parts)
422 995 aaronmk
funcs['_name'] = _name
423 102 aaronmk
424 2016 aaronmk
def _namePart(items, node):
425 328 aaronmk
    out_items = []
426
    for part, value in items:
427
        try: slice_ = name_parts_slices[part]
428 1612 aaronmk
        except KeyError, e: raise SyntaxError(e)
429 1219 aaronmk
        out_items.append((part, ' '.join(value.split(' ')[slice_])))
430 2016 aaronmk
    return _name(out_items, node)
431 995 aaronmk
funcs['_namePart'] = _namePart
432 1321 aaronmk
433 1607 aaronmk
#### Angles
434
435 2016 aaronmk
def _compass(items, node):
436 1607 aaronmk
    '''Converts a compass direction (N, NE, NNE, etc.) into a degree heading'''
437
    items = dict(conv_items(strings.ustr, items))
438
    try: value = items['value']
439
    except KeyError: return None # input is empty
440
441
    if not value.isupper(): return value # pass through other coordinate formats
442
    try: return util.cast(str, angles.compass2heading(value)) # ignore None
443
    except KeyError, e: raise FormatException(e)
444
funcs['_compass'] = _compass
445
446 1469 aaronmk
#### Paths
447
448 2016 aaronmk
def _simplifyPath(items, node):
449 1321 aaronmk
    items = dict(items)
450
    try:
451 1562 aaronmk
        next = cast(strings.ustr, items['next'])
452
        require = cast(strings.ustr, items['require'])
453 1321 aaronmk
        root = items['path']
454 1612 aaronmk
    except KeyError, e: raise SyntaxError(e)
455 1321 aaronmk
456
    node = root
457
    while node != None:
458
        new_node = xpath.get_1(node, next, allow_rooted=False)
459 2558 aaronmk
        if xpath.get_value(node, require, allow_rooted=False) == None: # empty
460 1321 aaronmk
            xml_dom.replace(node, new_node) # remove current elem
461
            if node is root: root = new_node # also update root
462
        node = new_node
463
    return root
464
funcs['_simplifyPath'] = _simplifyPath