-
Notifications
You must be signed in to change notification settings - Fork 2
/
oerp-pricelist.py
executable file
·518 lines (422 loc) · 17.5 KB
/
oerp-pricelist.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
#!/usr/bin/env python2.7
# -*- coding: utf-8 -*-
# (C) Max Gaukler, Julian Hammer 2014
# unlimited usage allowed, see LICENSE file
# Dependencies
from copy import deepcopy
import sys
import re
import oerplib
import locale
from ConfigParser import ConfigParser
import codecs
import cgi
import time
from repoze.lru import lru_cache
# LRU_CACHE_MAX_ENTRIES = 327678
LRU_CACHE_MAX_ENTRIES = None
import natsort
import json
import shutil
from tqdm import tqdm
from profiling import Profiler
import os
# switching to german:
locale.setlocale(locale.LC_ALL, "de_DE.UTF-8")
reload(sys).setdefaultencoding('UTF-8') # somehow doesn't work
if sys.stdout.encoding != "UTF-8":
print sys.stdout.encoding
print >> sys.stderr, "please use a UTF-8 locale, e.g. LANG=en_US.UTF-8"
exit(1)
cfg = ConfigParser({})
cfg.readfp(codecs.open('config.ini', 'r', 'utf8'))
oerp = oerplib.OERP(server=cfg.get('openerp', 'server'),
protocol='xmlrpc+ssl',
database=cfg.get('openerp', 'database'),
port=cfg.getint('openerp', 'port'),
version=cfg.get('openerp', 'version'))
user = oerp.login(user=cfg.get('openerp', 'user'),
passwd=cfg.get('openerp', 'password'))
oerpContext = oerp.context
from ratelimit import limits, sleep_and_retry
oerp_rl_calls = float(os.getenv('OERP_RATE_LIMIT_CALLS', '1'))
oerp_rl_period = float(os.getenv('OERP_RATE_LIMIT_PERIOD_SECONDS', '2'))
print("Rate limiting OERP API search calls to %s calls per %s seconds" % (oerp_rl_calls, oerp_rl_period))
@sleep_and_retry
@limits(calls=oerp_rl_calls,
period=oerp_rl_period)
def oerp_api_call(function, *args, **kwargs):
if function == 'search':
return oerp.search(*args, **kwargs)
else:
raise Exception("Unknown function '%s'" % function)
def oerp_search(*args, **kwargs):
return oerp_api_call('search', *args, **kwargs)
def oerp_read(*args, **kwargs):
return oerp.read(*args, **kwargs)
def oerp_write(*args, **kwargs):
return oerp.write(*args, **kwargs)
def oerp_create(*args, **kwargs):
return oerp.create(*args, **kwargs)
def str_to_int(s, fallback=None):
try:
return int(s)
except ValueError:
return fallback
@lru_cache(LRU_CACHE_MAX_ENTRIES)
def categ_id_to_list_of_names(c_id):
with Profiler("categ_id_to_list_of_names"):
# TODO make this faster by once fetching the list of all categories
categ = get_category(c_id)
if not categ['parent_id'] or \
categ['parent_id'][0] == cfg.getint('openerp', 'base_category_id'):
return [categ['name']]
else:
return categ_id_to_list_of_names(categ['parent_id'][0]) + [categ['name']]
class NotFound(Exception):
pass
def get_id(db, prod_filter):
ids = get_ids(db, prod_filter)
if not ids:
raise NotFound("cannot find {} from search {}".format(db, str(prod_filter)))
assert len(ids) == 1, "found more than one {} from search {}".format(db, str(prod_filter))
return ids[0]
def get_ids(db, prod_filter):
return oerp_search(db, prod_filter, context=oerpContext)
def read(db, prod_id, fields=None):
if not fields:
fields = []
assert type(
prod_id) == int, "read is only for one element. " \
"See also: read_elements() for reading multiple elements with a filter"
read_result = oerp_read(db, [prod_id], fields, context=oerpContext)
if len(read_result) != 1:
raise NotFound()
return read_result[0]
def write(db, prod_id, data):
return oerp_write(db, [prod_id], data, context=oerpContext)
def create(db, data):
return oerp_create(db, data, context=oerpContext)
def read_elements(db, element_filter, fields=None):
if not fields:
fields = []
return oerp_read(db, get_ids(db, element_filter), fields, context=oerpContext)
def read_property(db, prod_id, field, first_list_item=False):
read_result = read(db, prod_id, [field])
prod_property = read_result[field]
if first_list_item:
return prod_property[0]
else:
return prod_property
def category_id_from_name(name):
return get_id('product.category', [('name', '=', name)])
def get_category_with_descendants(cat_id):
"""
get IDs of all descendant categories (sub-category, sub-sub-category, ...),
including the original category itself
"""
return [cat_id] + get_category_descendants(cat_id)
@lru_cache(LRU_CACHE_MAX_ENTRIES)
def get_categories():
return read_elements('product.category', [], ['parent_id', 'name', 'property_stock_location'])
def get_category(cat_id):
for c in get_categories():
if c['id'] == cat_id:
return c
raise NotFound()
def get_category_children(cat_id):
# IDs of all direct child categories
for c in get_categories():
if c['parent_id'] and c['parent_id'][0] == cat_id:
yield c['id']
def get_category_descendants(cat_id):
"""
get IDs of all descendant categories (sub-category, sub-sub-category, ...),
excluding the original category itself
"""
children = list(get_category_children(cat_id))
descendants = list(children) # make a copy to avoid that 'children' is modified later
for x in children:
descendants += get_category_descendants(x)
return descendants
@lru_cache(LRU_CACHE_MAX_ENTRIES)
def get_supplier_info():
return read_elements('product.supplierinfo', [])
def get_supplier_info_from_product(p):
with Profiler("get_supplier_info_from_product"):
# input: product data dict
# TODO only shows first supplier
if len(p['seller_ids']) == 0:
raise NotFound
for i in get_supplier_info():
if i['id'] == p['seller_ids'][0]:
return i
raise NotFound()
@lru_cache(LRU_CACHE_MAX_ENTRIES)
def fetch_stock_location(location_id):
return oerp_read('stock.location', location_id)
@lru_cache(LRU_CACHE_MAX_ENTRIES)
def fetch_product_uom(uom_id):
return oerp_read('product.uom', uom_id)
def get_location_str_from_product(p):
with Profiler("get_location_str_from_product"):
location = p['property_stock_location']
if not location:
# no location given for product
# fall back to category's location like OERP storage module does
c = get_category(p['categ_id'][0])
location = c['property_stock_location']
if location:
location_id = location[0]
location_string = location[1]
location = fetch_stock_location(location_id)
if location['code']:
location_string += u" ({})".format(location['code'])
for removePrefix in [u"tats\xe4chliche Lagerorte / FAU FabLab / ",
u"tats\xe4chliche Lagerorte / "]:
if location_string.startswith(removePrefix):
location_string = location_string[len(removePrefix):]
else:
# no location set at all
location_string = "kein Ort eingetragen"
return location_string
def get_uom_rounding_from_product(p):
with Profiler("get_uom_rounding_from_product"):
uom_id = p['uom_id']
rounding = 1.0
if uom_id:
product_uom = fetch_product_uom(uom_id)
if product_uom['rounding']:
rounding = product_uom['rounding']
else:
# no rounding set at all
# assume rounding 1
rounding = 1.0
return rounding
def _parse_product(p):
"""
takes a product from oerp as a dictionary,
re-formats values and adds calculated ones
"""
# product code (article number, PLU)
p['_code_str'] = "{:04d}".format(int(p['code']))
# category as list ["A","B","foo"]
p['_categ_list'] = categ_id_to_list_of_names(p['categ_id'][0])
# _categ_str: category as one string "A / B / foo"
p['_categ_str'] = " / ".join(p['_categ_list'])
# _price_str: list price as string
price_str = '{:.3f}'.format(p['lst_price']).replace(".", ",")
if price_str[-1] == "0": # third digit only if nonzero
price_str = price_str[:-1]
p['_price_str'] = u'{} €'.format(price_str)
if p['lst_price'] == 0:
p['_price_str'] = u"gegen Spende"
if not p['sale_ok']:
p['_price_str'] = u"unverkäuflich"
p['_name_and_description'] = p['name']
if p['description']:
p['_name_and_description'] += '\n' + p['description']
# _supplier_all_infos; supplier and manufacturer info
p['_supplier_all_infos'] = ''
try:
p['_supplierinfo'] = get_supplier_info_from_product(p)
if p['_supplierinfo']['name']:
p['_supplier_name'] = p['_supplierinfo']['name'][1]
p['_supplier_code'] = p['_supplierinfo']['product_code'] or ''
p['_supplier_name_code'] = '{}: {}'.format(
p['_supplier_name'], p['_supplier_code'])
p['_supplier_all_infos'] += p['_supplier_name_code']
except NotFound:
pass
if p['manufacturer']:
p['_supplier_all_infos'] += ", Hersteller: {} ".format(p['manufacturer'][1])
if p['manufacturer_pref']:
p['_supplier_all_infos'] += "({}) ".format(p['manufacturer_pref'])
p['_supplier_all_infos'] += p['manufacturer_pname'] or ''
p['_uom_str'] = p['uom_id'][1]
p['uom_id'] = p['uom_id'][0]
p['_location_str'] = get_location_str_from_product(p)
p['_uom_rounding'] = get_uom_rounding_from_product(p)
# String für Verkaufseinheit: "pro Stück"
if p['lst_price'] == 0 or p["_uom_str"].startswith("bei"):
# keine Einheit anzeigen, wenn
# - Preis null
# oder
# - Einheit mit "bei" anfängt, weil "pro bei" Unfug ist
p['_per_uom_str'] = p['_uom_str']
else:
p['_per_uom_str'] = "pro " + p['_uom_str']
return p
def import_products_oerp(cat_name, data, extra_filters=None, columns=None):
# TODO code vs default_code -> what's the difference?
if not columns:
columns = []
if not extra_filters:
extra_filters = []
columns = deepcopy(columns)
if columns:
columns += ["name",
"description",
"code", "default_code",
"lst_price",
"active",
"sale_ok",
"categ_id",
"uom_id",
"manufacturer",
"manufacturer_pname",
"manufacturer_pref",
"seller_ids",
"property_stock_location"]
prod_ids = oerp_search('product.product', [('default_code', '!=', False)] + extra_filters)
prods = []
def split_list(prod_list, chunk_size):
return [prod_list[offs:offs + chunk_size] for offs in range(0, len(prod_list), chunk_size)]
# columns starting with _ are generated in this script and not from the DB
query_columns = [col for col in columns if not col.startswith("_")]
chunk_size = 100
with tqdm(total=len(prod_ids), desc="Fetching products of category '{}'".format(cat_name), leave=False) as pbar:
for prod_ids_slice in split_list(prod_ids, chunk_size):
prods += oerp_read('product.product', prod_ids_slice, query_columns, context=oerp.context)
pbar.update(len(prod_ids_slice))
# Only consider things with numerical PLUs in code field
prods = filter(lambda p: str_to_int(p['code']) is not None, prods)
for p in tqdm(prods, desc="Processing products of category '{}'".format(cat_name), leave=False):
if not p['active'] or not p['sale_ok']:
continue
p = _parse_product(p)
data[p['code']] = p
return data
def html_escape(x):
return cgi.escape(x).encode('ascii', 'xmlcharrefreplace').replace("\n", "<br/>")
def tr(x, tr_options="", td_options=None, escape=True):
"""
creates a html table row (<tr>) out of the dict x.
If you want to provide html properties for tr, use tr_options
If you want to provide html properties for the tds, use td_options
If escape==False, the values of x won't be html escaped
"""
out = u"<tr {}>".format(tr_options)
for v in x:
out += u"<td {}>{}</td>".format(
td_options[x.index(v)] if td_options else "",
html_escape(v) if escape else v)
out += u"</tr>"
return out
# fill the template file with data
def make_html_from_template(heading, content):
with open("template.html", "r") as f:
out = f.read()
f.close()
out = out.replace("$HEADING", heading)
out = out.replace("$REFRESHDATE", time.strftime("%x %X", time.localtime()))
out = out.replace("$CONTENTTABLE", content)
return out
def make_price_list_html(base_category, columns, column_names):
if type(base_category) != int:
base_category = category_id_from_name(base_category)
categories = get_category_with_descendants(base_category)
data = import_products_oerp(base_category, {}, [('categ_id', 'in', categories)], columns)
jsondata = json.dumps(data)
def make_header(x):
return column_names.get(x, x)
content_table = tr([make_header(x) for x in columns], 'class="head"')
product_list = data.values()
product_list = natsort.natsorted(product_list, key=lambda x: [x['_categ_str'], x['name']])
current_category = None
for p in product_list:
if p['_categ_str'] != current_category:
# Make a heading for the new category the current product belongs
current_category = p['_categ_str']
content_table += u'''
<tr id="{categ_str}" class="newCateg">
<td colspan="{colspan}">
<a id="permalink" href="#{categ_str}" title="Permalink">¶</a>
{categ_str}
</td>
</tr>
'''.format(
categ_str=html_escape(p["_categ_str"]),
colspan=len(columns))
row = []
for w in columns:
if w == '_code_str':
# add the permalink ¶
row.append(u'''
<a id="permalink" href="#{default_code}" title="Permalink">¶</a>
{default_code}
'''.format(
default_code=html_escape(p['default_code'])
))
else:
# escape, as tr musn't escape because of the permalink <a>
row.append(html_escape(str(p.get(w, ""))))
td_props = [''] * len(columns)
td_props[0] = 'style="width:50px;"'
content_table += tr(row,
tr_options='id="{}"'.format(p['default_code']),
td_options=td_props,
escape=False)
heading = "Preisliste " + " / ".join(categ_id_to_list_of_names(base_category))
out = make_html_from_template(heading, content_table)
return heading, out, jsondata
def main():
column_names = {"_code_str": "Nr.",
"_name_and_description": "Bezeichnung",
"_price_str": "Preis",
"_uom_str": "Einheit",
"_supplier_name_code": "Lieferant",
"_supplier_all_infos": "Lieferant / Hersteller",
"_location_str": "Ort"}
# examples for custom columns, added via an attribute set in ERP:
# "x_durchmesser": "D", "x_stirnseitig": "eintauchen?", "x_fraeserwerkstoff": "aus Material",
default_cols = ["_code_str", "_name_and_description", "_price_str",
"_uom_str", "_location_str", "_supplier_all_infos"]
jobs = [ # ("Fräser", default_cols+["x_durchmesser", "x_stirnseitig",
# "x_fraeserwerkstoff", "x_fuerwerkstoff"]),
("CNC", default_cols),
(228, default_cols), # Fräsenmaterial
("Laser", default_cols),
("Schneideplotter", default_cols),
("Textilbearbeitung", default_cols),
("Platinenfertigung", default_cols),
("Alle Produkte", default_cols)
]
file_list = []
for (cat, columns) in tqdm(jobs, desc="Generating HTML export of OERP data"):
(title, price_list, jsondata) = make_price_list_html(cat, columns, column_names)
if type(cat) == int:
cat = str(cat)
filename = "price_list-{}.html".format(re.sub(r'[^0-9a-zA-Z]', '_', cat))
f = open("output/" + filename, "w")
f.write(price_list)
file_list.append((filename, title))
f.close()
f = open("output/" + filename + ".json", "w");
f.write(jsondata)
f.close()
with open("output/categories_Alle_Produkte.json", "w") as f:
# generate a json of categories for category "Alle Produkte"
base_category = "Alle Produkte"
if type(base_category) != int:
base_category = category_id_from_name(base_category)
categories = get_category_with_descendants(base_category)
out = []
for category in categories:
out.append(get_category(category))
jsondata = json.dumps(out)
f.write(jsondata)
shutil.copyfile("shop.html", "output/shop.html")
f = open("output/index.html", "w")
html_list = "<ul>"
file_list = [("shop.html", "Preisrechner Alle Produkte")] + file_list
for (filename, title) in file_list:
html_list += '<li><a href="{}">{}</a></li>'.format(filename, html_escape(title))
html_list += "</ul>"
html_list = make_html_from_template("Übersicht aller Preislisten", html_list)
f.write(html_list)
f.close()
# print("Execution stats:")
# print(Profiler.stats)
if __name__ == '__main__':
main()