1
|
--
|
2
|
-- PostgreSQL database dump
|
3
|
--
|
4
|
|
5
|
SET statement_timeout = 0;
|
6
|
SET client_encoding = 'UTF8';
|
7
|
SET standard_conforming_strings = on;
|
8
|
SET check_function_bodies = false;
|
9
|
SET client_min_messages = warning;
|
10
|
|
11
|
--
|
12
|
-- Name: py_util; Type: SCHEMA; Schema: -; Owner: bien
|
13
|
--
|
14
|
|
15
|
CREATE SCHEMA py_util;
|
16
|
|
17
|
|
18
|
ALTER SCHEMA py_util OWNER TO bien;
|
19
|
|
20
|
--
|
21
|
-- Name: SCHEMA py_util; Type: COMMENT; Schema: -; Owner: bien
|
22
|
--
|
23
|
|
24
|
COMMENT ON SCHEMA py_util IS 'IMPORTANT: Functions must always return NULL in place of '''' (the empty string). This ensures that empty strings do not find their way into VegBIEN.';
|
25
|
|
26
|
|
27
|
SET search_path = py_util, pg_catalog;
|
28
|
|
29
|
--
|
30
|
-- Name: _date(timestamp with time zone); Type: FUNCTION; Schema: py_util; Owner: bien
|
31
|
--
|
32
|
|
33
|
CREATE FUNCTION _date(date timestamp with time zone) RETURNS timestamp with time zone
|
34
|
LANGUAGE sql IMMUTABLE STRICT
|
35
|
AS $_$
|
36
|
SELECT $1
|
37
|
$_$;
|
38
|
|
39
|
|
40
|
ALTER FUNCTION py_util._date(date timestamp with time zone) OWNER TO bien;
|
41
|
|
42
|
--
|
43
|
-- Name: _date(text, text, text); Type: FUNCTION; Schema: py_util; Owner: bien
|
44
|
--
|
45
|
|
46
|
CREATE FUNCTION _date(year text DEFAULT NULL::text, month text DEFAULT NULL::text, day text DEFAULT NULL::text) RETURNS text
|
47
|
LANGUAGE plpython3u IMMUTABLE
|
48
|
AS $$
|
49
|
global date, year, month, day
|
50
|
|
51
|
import datetime
|
52
|
|
53
|
def e_msg(e): return e.args[0].rstrip()
|
54
|
|
55
|
# Year is required
|
56
|
if year == None:
|
57
|
if month == None and day == None: return None # entire date is empty
|
58
|
else: raise AssertionError(
|
59
|
'null value in column "year" violates not-null constraint')
|
60
|
|
61
|
# Convert month name to number
|
62
|
if month != None and not month.isdigit(): # month is name
|
63
|
month = str(datetime.datetime.strptime(month, '%b').month)
|
64
|
|
65
|
if month == None: month = 1
|
66
|
if day == None: day = 1
|
67
|
year, month, day = map(int, (year, month, day))
|
68
|
|
69
|
for try_num in range(2):
|
70
|
try:
|
71
|
date = datetime.date(year, month, day)
|
72
|
break
|
73
|
except ValueError as e:
|
74
|
if try_num > 0: raise # exception still raised after retry
|
75
|
msg = e_msg(e)
|
76
|
if msg == 'month must be in 1..12': # try swapping month and day
|
77
|
month, day = day, month
|
78
|
else: raise
|
79
|
|
80
|
return str(date)
|
81
|
$$;
|
82
|
|
83
|
|
84
|
ALTER FUNCTION py_util._date(year text, month text, day text) OWNER TO bien;
|
85
|
|
86
|
--
|
87
|
-- Name: _date(integer, integer, integer); Type: FUNCTION; Schema: py_util; Owner: bien
|
88
|
--
|
89
|
|
90
|
CREATE FUNCTION _date(year integer DEFAULT NULL::integer, month integer DEFAULT NULL::integer, day integer DEFAULT NULL::integer) RETURNS text
|
91
|
LANGUAGE sql IMMUTABLE STRICT
|
92
|
AS $_$
|
93
|
SELECT py_util._date($1::text, $2::text, $3::text)
|
94
|
$_$;
|
95
|
|
96
|
|
97
|
ALTER FUNCTION py_util._date(year integer, month integer, day integer) OWNER TO bien;
|
98
|
|
99
|
--
|
100
|
-- Name: _dateRangeEnd(text); Type: FUNCTION; Schema: py_util; Owner: bien
|
101
|
--
|
102
|
|
103
|
CREATE FUNCTION "_dateRangeEnd"(value text) RETURNS text
|
104
|
LANGUAGE sql IMMUTABLE STRICT
|
105
|
AS $_$
|
106
|
SELECT (py_util.parse_date_range($1))[2]
|
107
|
$_$;
|
108
|
|
109
|
|
110
|
ALTER FUNCTION py_util."_dateRangeEnd"(value text) OWNER TO bien;
|
111
|
|
112
|
--
|
113
|
-- Name: _dateRangeEnd(timestamp with time zone); Type: FUNCTION; Schema: py_util; Owner: bien
|
114
|
--
|
115
|
|
116
|
CREATE FUNCTION "_dateRangeEnd"(value timestamp with time zone) RETURNS timestamp with time zone
|
117
|
LANGUAGE sql IMMUTABLE STRICT
|
118
|
AS $_$
|
119
|
SELECT $1
|
120
|
$_$;
|
121
|
|
122
|
|
123
|
ALTER FUNCTION py_util."_dateRangeEnd"(value timestamp with time zone) OWNER TO bien;
|
124
|
|
125
|
--
|
126
|
-- Name: _dateRangeStart(text); Type: FUNCTION; Schema: py_util; Owner: bien
|
127
|
--
|
128
|
|
129
|
CREATE FUNCTION "_dateRangeStart"(value text) RETURNS text
|
130
|
LANGUAGE sql IMMUTABLE STRICT
|
131
|
AS $_$
|
132
|
SELECT (py_util.parse_date_range($1))[1]
|
133
|
$_$;
|
134
|
|
135
|
|
136
|
ALTER FUNCTION py_util."_dateRangeStart"(value text) OWNER TO bien;
|
137
|
|
138
|
--
|
139
|
-- Name: _dateRangeStart(timestamp with time zone); Type: FUNCTION; Schema: py_util; Owner: bien
|
140
|
--
|
141
|
|
142
|
CREATE FUNCTION "_dateRangeStart"(value timestamp with time zone) RETURNS timestamp with time zone
|
143
|
LANGUAGE sql IMMUTABLE STRICT
|
144
|
AS $_$
|
145
|
SELECT $1
|
146
|
$_$;
|
147
|
|
148
|
|
149
|
ALTER FUNCTION py_util."_dateRangeStart"(value timestamp with time zone) OWNER TO bien;
|
150
|
|
151
|
--
|
152
|
-- Name: _namePart(text, text, text); Type: FUNCTION; Schema: py_util; Owner: bien
|
153
|
--
|
154
|
|
155
|
CREATE FUNCTION "_namePart"(first text DEFAULT NULL::text, middle text DEFAULT NULL::text, last text DEFAULT NULL::text) RETURNS text
|
156
|
LANGUAGE plpython3u IMMUTABLE
|
157
|
AS $$
|
158
|
params = dict(first=first, middle=middle, last=last)
|
159
|
|
160
|
def none_if(val, *none_vals):
|
161
|
for none_val in none_vals:
|
162
|
if val == none_val: return None
|
163
|
return val
|
164
|
|
165
|
_name_parts_slices_items = [
|
166
|
('first', slice(None, -1)),
|
167
|
('middle', slice(1, -1)),
|
168
|
('last', slice(-1, None)),
|
169
|
]
|
170
|
name_parts_slices = dict(_name_parts_slices_items)
|
171
|
name_parts = [name for name, slice_ in _name_parts_slices_items]
|
172
|
|
173
|
def _name(items):
|
174
|
items = dict(items)
|
175
|
parts = []
|
176
|
for part in name_parts:
|
177
|
if part in items: parts.append(items[part])
|
178
|
return none_if(' '.join(parts), '')
|
179
|
|
180
|
out_items = []
|
181
|
for part, value in params.iteritems():
|
182
|
if value == None: continue
|
183
|
|
184
|
try: slice_ = name_parts_slices[part]
|
185
|
except KeyError: pass # a non-value column
|
186
|
else: out_items.append((part, ' '.join(value.split(' ')[slice_])))
|
187
|
|
188
|
return _name(out_items)
|
189
|
$$;
|
190
|
|
191
|
|
192
|
ALTER FUNCTION py_util."_namePart"(first text, middle text, last text) OWNER TO bien;
|
193
|
|
194
|
--
|
195
|
-- Name: parse_date_range(text); Type: FUNCTION; Schema: py_util; Owner: bien
|
196
|
--
|
197
|
|
198
|
CREATE FUNCTION parse_date_range(str_ text) RETURNS text[]
|
199
|
LANGUAGE plpython3u IMMUTABLE STRICT
|
200
|
AS $$
|
201
|
import re
|
202
|
|
203
|
def single_space(str_): return re.sub(r' {2,}', r' ', str_.strip())
|
204
|
|
205
|
def could_be_year(str_): return str_.isdigit() and len(str_) == 4
|
206
|
|
207
|
def could_be_day(str_): return str_.isdigit() and len(str_) <= 2
|
208
|
|
209
|
range_sep='-'
|
210
|
part_sep=' '
|
211
|
|
212
|
default = (str_, str_)
|
213
|
if str_.find(':') >= 0: return default
|
214
|
# range_sep might be used as date part separator instead
|
215
|
if str_.find(part_sep) < 0: return default
|
216
|
|
217
|
start, sep, end = str_.partition(range_sep)
|
218
|
if sep == '': return default # not a range
|
219
|
start, end = (single_space(d).split(part_sep) for d in (start, end))
|
220
|
|
221
|
# Has form M D1-D2 or M D1-D2 Y (not M1 Y1-M2 Y2 or M1 D1-M2 D2)
|
222
|
if len(start) == 2 and (len(end) == 1 or (
|
223
|
len(end) == 2 and could_be_day(start[-1]) and could_be_day(end[0])
|
224
|
and could_be_year(end[-1])
|
225
|
)):
|
226
|
end.insert(0, start[0]) # make end fully specified
|
227
|
ct_diff = len(end) - len(start)
|
228
|
# Has form D1-D2 M Y, M1 D1-M2 D2 Y, M1-M2 Y, etc.
|
229
|
if ct_diff > 0: start += end[-ct_diff:] # make start fully specified
|
230
|
# Other forms are invalid and will be left as-is
|
231
|
|
232
|
return [part_sep.join(d) for d in (start, end)]
|
233
|
$$;
|
234
|
|
235
|
|
236
|
ALTER FUNCTION py_util.parse_date_range(str_ text) OWNER TO bien;
|
237
|
|
238
|
--
|
239
|
-- PostgreSQL database dump complete
|
240
|
--
|
241
|
|