96884afeb6272390af65c98a3a143d2e2be5c972
[dmarc.git] / write-to-database.ipynb
1 {
2 "cells": [
3 {
4 "cell_type": "code",
5 "execution_count": 92,
6 "metadata": {
7 "collapsed": true
8 },
9 "outputs": [],
10 "source": [
11 "import configparser\n",
12 "import xml.etree.ElementTree\n",
13 "import psycopg2\n",
14 "import re\n",
15 "import datetime"
16 ]
17 },
18 {
19 "cell_type": "code",
20 "execution_count": 2,
21 "metadata": {
22 "collapsed": false
23 },
24 "outputs": [
25 {
26 "data": {
27 "text/plain": [
28 "['dmarc.ini']"
29 ]
30 },
31 "execution_count": 2,
32 "metadata": {},
33 "output_type": "execute_result"
34 }
35 ],
36 "source": [
37 "config = configparser.ConfigParser()\n",
38 "config.read('dmarc.ini')"
39 ]
40 },
41 {
42 "cell_type": "code",
43 "execution_count": 4,
44 "metadata": {
45 "collapsed": false
46 },
47 "outputs": [
48 {
49 "data": {
50 "text/plain": [
51 "['imap', 'database']"
52 ]
53 },
54 "execution_count": 4,
55 "metadata": {},
56 "output_type": "execute_result"
57 }
58 ],
59 "source": [
60 "config.sections()"
61 ]
62 },
63 {
64 "cell_type": "code",
65 "execution_count": 6,
66 "metadata": {
67 "collapsed": false
68 },
69 "outputs": [
70 {
71 "data": {
72 "text/plain": [
73 "['server', 'username', 'port', 'password']"
74 ]
75 },
76 "execution_count": 6,
77 "metadata": {},
78 "output_type": "execute_result"
79 }
80 ],
81 "source": [
82 "[k for k in config['imap']]"
83 ]
84 },
85 {
86 "cell_type": "code",
87 "execution_count": 7,
88 "metadata": {
89 "collapsed": false
90 },
91 "outputs": [
92 {
93 "data": {
94 "text/plain": [
95 "['server', 'database', 'username', 'password']"
96 ]
97 },
98 "execution_count": 7,
99 "metadata": {},
100 "output_type": "execute_result"
101 }
102 ],
103 "source": [
104 "[k for k in config['database']]"
105 ]
106 },
107 {
108 "cell_type": "code",
109 "execution_count": 25,
110 "metadata": {
111 "collapsed": true
112 },
113 "outputs": [],
114 "source": [
115 "def walk(node, indent=0):\n",
116 " for child in node:\n",
117 " if child.text:\n",
118 " txt = child.text.strip()\n",
119 " else:\n",
120 " txt = ''\n",
121 " print(' '*indent, child.tag, ':', txt)\n",
122 " walk(child, indent+2)"
123 ]
124 },
125 {
126 "cell_type": "code",
127 "execution_count": 16,
128 "metadata": {
129 "collapsed": false
130 },
131 "outputs": [
132 {
133 "name": "stdout",
134 "output_type": "stream",
135 "text": [
136 "google.com!njae.me.uk!1458777600!1458863999.xml hotmail.com!njae.me.uk!1459011600!1459098000.xml yahoo.com!njae.me.uk!1459036800!1459123199.xml\r\n",
137 "google.com!njae.me.uk!1458950400!1459036799.xml yahoo.com!njae.me.uk!1458864000!1458950399.xml\r\n"
138 ]
139 }
140 ],
141 "source": [
142 "!ls *xml"
143 ]
144 },
145 {
146 "cell_type": "code",
147 "execution_count": 17,
148 "metadata": {
149 "collapsed": false
150 },
151 "outputs": [
152 {
153 "data": {
154 "text/plain": [
155 "['google.com!njae.me.uk!1458777600!1458863999.xml',\n",
156 " 'google.com!njae.me.uk!1458950400!1459036799.xml',\n",
157 " 'hotmail.com!njae.me.uk!1459011600!1459098000.xml',\n",
158 " 'yahoo.com!njae.me.uk!1458864000!1458950399.xml',\n",
159 " 'yahoo.com!njae.me.uk!1459036800!1459123199.xml']"
160 ]
161 },
162 "execution_count": 17,
163 "metadata": {},
164 "output_type": "execute_result"
165 }
166 ],
167 "source": [
168 "xmls = !ls *xml\n",
169 "xmls"
170 ]
171 },
172 {
173 "cell_type": "code",
174 "execution_count": 19,
175 "metadata": {
176 "collapsed": false
177 },
178 "outputs": [
179 {
180 "data": {
181 "text/plain": [
182 "{'google.com!njae.me.uk!1458777600!1458863999.xml': <Element 'feedback' at 0x7fa9a8169a98>,\n",
183 " 'google.com!njae.me.uk!1458950400!1459036799.xml': <Element 'feedback' at 0x7fa9a8170a48>,\n",
184 " 'hotmail.com!njae.me.uk!1459011600!1459098000.xml': <Element 'feedback' at 0x7fa9a80f4a48>,\n",
185 " 'yahoo.com!njae.me.uk!1458864000!1458950399.xml': <Element 'feedback' at 0x7fa9a80f74a8>,\n",
186 " 'yahoo.com!njae.me.uk!1459036800!1459123199.xml': <Element 'feedback' at 0x7fa9a80f7ea8>}"
187 ]
188 },
189 "execution_count": 19,
190 "metadata": {},
191 "output_type": "execute_result"
192 }
193 ],
194 "source": [
195 "dmarc_reports = {f: xml.etree.ElementTree.fromstring(open(f).read()) for f in xmls}\n",
196 "dmarc_reports"
197 ]
198 },
199 {
200 "cell_type": "code",
201 "execution_count": 26,
202 "metadata": {
203 "collapsed": false
204 },
205 "outputs": [
206 {
207 "name": "stdout",
208 "output_type": "stream",
209 "text": [
210 " report_metadata : \n",
211 " org_name : google.com\n",
212 " email : noreply-dmarc-support@google.com\n",
213 " extra_contact_info : https://support.google.com/a/answer/2466580\n",
214 " report_id : 2150510829392606201\n",
215 " date_range : \n",
216 " begin : 1458950400\n",
217 " end : 1459036799\n",
218 " policy_published : \n",
219 " domain : njae.me.uk\n",
220 " adkim : r\n",
221 " aspf : r\n",
222 " p : none\n",
223 " sp : none\n",
224 " pct : 100\n",
225 " record : \n",
226 " row : \n",
227 " source_ip : 82.109.184.9\n",
228 " count : 1\n",
229 " policy_evaluated : \n",
230 " disposition : none\n",
231 " dkim : fail\n",
232 " spf : fail\n",
233 " identifiers : \n",
234 " header_from : njae.me.uk\n",
235 " auth_results : \n",
236 " dkim : \n",
237 " domain : clublloyds.com\n",
238 " result : pass\n",
239 " spf : \n",
240 " domain : clublloyds.com\n",
241 " result : pass\n",
242 " record : \n",
243 " row : \n",
244 " source_ip : 212.69.55.62\n",
245 " count : 1\n",
246 " policy_evaluated : \n",
247 " disposition : none\n",
248 " dkim : pass\n",
249 " spf : pass\n",
250 " identifiers : \n",
251 " header_from : njae.me.uk\n",
252 " auth_results : \n",
253 " dkim : \n",
254 " domain : njae.me.uk\n",
255 " result : pass\n",
256 " spf : \n",
257 " domain : njae.me.uk\n",
258 " result : pass\n",
259 "\n",
260 " report_metadata : \n",
261 " org_name : Yahoo! Inc.\n",
262 " email : postmaster@dmarc.yahoo.com\n",
263 " report_id : 1459129809.695034\n",
264 " date_range : \n",
265 " begin : 1459036800\n",
266 " end : 1459123199\n",
267 " policy_published : \n",
268 " domain : njae.me.uk\n",
269 " adkim : r\n",
270 " aspf : r\n",
271 " p : none\n",
272 " pct : 100\n",
273 " record : \n",
274 " row : \n",
275 " source_ip : 65.20.0.12\n",
276 " count : 1\n",
277 " policy_evaluated : \n",
278 " disposition : none\n",
279 " dkim : pass\n",
280 " spf : fail\n",
281 " identifiers : \n",
282 " header_from : njae.me.uk\n",
283 " auth_results : \n",
284 " dkim : \n",
285 " domain : njae.me.uk\n",
286 " result : pass\n",
287 " spf : \n",
288 " domain : njae.me.uk\n",
289 " result : softfail\n",
290 "\n",
291 " report_metadata : \n",
292 " org_name : Yahoo! Inc.\n",
293 " email : postmaster@dmarc.yahoo.com\n",
294 " report_id : 1458957186.548175\n",
295 " date_range : \n",
296 " begin : 1458864000\n",
297 " end : 1458950399\n",
298 " policy_published : \n",
299 " domain : njae.me.uk\n",
300 " adkim : r\n",
301 " aspf : r\n",
302 " p : none\n",
303 " pct : 100\n",
304 " record : \n",
305 " row : \n",
306 " source_ip : 212.69.55.62\n",
307 " count : 2\n",
308 " policy_evaluated : \n",
309 " disposition : none\n",
310 " dkim : pass\n",
311 " spf : pass\n",
312 " identifiers : \n",
313 " header_from : njae.me.uk\n",
314 " auth_results : \n",
315 " dkim : \n",
316 " domain : njae.me.uk\n",
317 " result : pass\n",
318 " spf : \n",
319 " domain : njae.me.uk\n",
320 " result : pass\n",
321 "\n",
322 " report_metadata : \n",
323 " org_name : Microsoft Corp.\n",
324 " email : dmarcrep@microsoft.com\n",
325 " report_id : 68aad5080a774e2c997d159b546569b9@hotmail.com\n",
326 " date_range : \n",
327 " begin : 1459011600\n",
328 " end : 1459098000\n",
329 " policy_published : \n",
330 " domain : njae.me.uk\n",
331 " adkim : r\n",
332 " aspf : r\n",
333 " p : none\n",
334 " sp : none\n",
335 " pct : 100\n",
336 " record : \n",
337 " row : \n",
338 " source_ip : 212.69.55.62\n",
339 " count : 2\n",
340 " policy_evaluated : \n",
341 " disposition : none\n",
342 " dkim : pass\n",
343 " spf : pass\n",
344 " identifiers : \n",
345 " header_from : njae.me.uk\n",
346 " auth_results : \n",
347 " spf : \n",
348 " domain : njae.me.uk\n",
349 " result : pass\n",
350 " dkim : \n",
351 " domain : njae.me.uk\n",
352 " result : pass\n",
353 "\n",
354 " report_metadata : \n",
355 " org_name : google.com\n",
356 " email : noreply-dmarc-support@google.com\n",
357 " extra_contact_info : https://support.google.com/a/answer/2466580\n",
358 " report_id : 13032699446207263761\n",
359 " date_range : \n",
360 " begin : 1458777600\n",
361 " end : 1458863999\n",
362 " policy_published : \n",
363 " domain : njae.me.uk\n",
364 " adkim : r\n",
365 " aspf : r\n",
366 " p : none\n",
367 " sp : none\n",
368 " pct : 100\n",
369 " record : \n",
370 " row : \n",
371 " source_ip : 146.185.136.235\n",
372 " count : 1\n",
373 " policy_evaluated : \n",
374 " disposition : none\n",
375 " dkim : pass\n",
376 " spf : fail\n",
377 " identifiers : \n",
378 " header_from : njae.me.uk\n",
379 " auth_results : \n",
380 " dkim : \n",
381 " domain : njae.me.uk\n",
382 " result : pass\n",
383 " spf : \n",
384 " domain : njae.me.uk\n",
385 " result : softfail\n",
386 "\n"
387 ]
388 }
389 ],
390 "source": [
391 "for f in dmarc_reports:\n",
392 " walk(dmarc_reports[f])\n",
393 " print()"
394 ]
395 },
396 {
397 "cell_type": "code",
398 "execution_count": 15,
399 "metadata": {
400 "collapsed": false
401 },
402 "outputs": [
403 {
404 "name": "stdout",
405 "output_type": "stream",
406 "text": [
407 " report_metadata : \n",
408 " org_name : Yahoo! Inc.\n",
409 " email : postmaster@dmarc.yahoo.com\n",
410 " report_id : 1459129809.695034\n",
411 " date_range : \n",
412 " begin : 1459036800\n",
413 " end : 1459123199\n",
414 " policy_published : \n",
415 " domain : njae.me.uk\n",
416 " adkim : r\n",
417 " aspf : r\n",
418 " p : none\n",
419 " pct : 100\n",
420 " record : \n",
421 " row : \n",
422 " source_ip : 65.20.0.12\n",
423 " count : 1\n",
424 " policy_evaluated : \n",
425 " disposition : none\n",
426 " dkim : pass\n",
427 " spf : fail\n",
428 " identifiers : \n",
429 " header_from : njae.me.uk\n",
430 " auth_results : \n",
431 " dkim : \n",
432 " domain : njae.me.uk\n",
433 " result : pass\n",
434 " spf : \n",
435 " domain : njae.me.uk\n",
436 " result : softfail\n"
437 ]
438 }
439 ],
440 "source": [
441 "root2 = xml.etree.ElementTree.fromstring(open('yahoo.com!njae.me.uk!1459036800!1459123199.xml').read())\n",
442 "walk(root2)"
443 ]
444 },
445 {
446 "cell_type": "code",
447 "execution_count": 28,
448 "metadata": {
449 "collapsed": false
450 },
451 "outputs": [],
452 "source": [
453 "conn = psycopg2.connect(host=config['database']['server'],\n",
454 " database=config['database']['database'], \n",
455 " user=config['database']['username'], \n",
456 " password=config['database']['password'])"
457 ]
458 },
459 {
460 "cell_type": "code",
461 "execution_count": 30,
462 "metadata": {
463 "collapsed": false
464 },
465 "outputs": [
466 {
467 "data": {
468 "text/plain": [
469 "<cursor object at 0x7fa9a819b7c8; closed: 0>"
470 ]
471 },
472 "execution_count": 30,
473 "metadata": {},
474 "output_type": "execute_result"
475 }
476 ],
477 "source": [
478 "cur = conn.cursor()\n",
479 "cur"
480 ]
481 },
482 {
483 "cell_type": "code",
484 "execution_count": 31,
485 "metadata": {
486 "collapsed": false
487 },
488 "outputs": [
489 {
490 "data": {
491 "text/plain": [
492 "{'email': 'varchar', 'org_name': 'varchar', 'report_id': 'varchar'}"
493 ]
494 },
495 "execution_count": 31,
496 "metadata": {},
497 "output_type": "execute_result"
498 }
499 ],
500 "source": [
501 "report_fields = {'date_range_begin': 'timestamp', 'date_range_end': 'timestamp'}\n",
502 "for child in root2:\n",
503 " if child.tag == 'report_metadata':\n",
504 " for c in child:\n",
505 " if c.tag != 'date_range':\n",
506 " report_fields[c.tag] = 'varchar'\n",
507 "report_fields"
508 ]
509 },
510 {
511 "cell_type": "code",
512 "execution_count": 65,
513 "metadata": {
514 "collapsed": false
515 },
516 "outputs": [],
517 "source": [
518 "def xpath_of(node, prefix=''):\n",
519 " if len(node) == 0:\n",
520 " if node.text:\n",
521 " nodes = {prefix + node.tag: node.text.strip()}\n",
522 " else:\n",
523 " nodes = {prefix + node.tag: ''}\n",
524 " else:\n",
525 " nodes = {}\n",
526 " if node:\n",
527 " for child in node:\n",
528 " nodes.update(xpath_of(child, \n",
529 " prefix + node.tag + '/'))\n",
530 " return nodes"
531 ]
532 },
533 {
534 "cell_type": "code",
535 "execution_count": 66,
536 "metadata": {
537 "collapsed": false
538 },
539 "outputs": [
540 {
541 "data": {
542 "text/plain": [
543 "{'feedback/policy_published/adkim': 'r',\n",
544 " 'feedback/policy_published/aspf': 'r',\n",
545 " 'feedback/policy_published/domain': 'njae.me.uk',\n",
546 " 'feedback/policy_published/p': 'none',\n",
547 " 'feedback/policy_published/pct': '100',\n",
548 " 'feedback/record/auth_results/dkim/domain': 'njae.me.uk',\n",
549 " 'feedback/record/auth_results/dkim/result': 'pass',\n",
550 " 'feedback/record/auth_results/spf/domain': 'njae.me.uk',\n",
551 " 'feedback/record/auth_results/spf/result': 'softfail',\n",
552 " 'feedback/record/identifiers/header_from': 'njae.me.uk',\n",
553 " 'feedback/record/row/count': '1',\n",
554 " 'feedback/record/row/policy_evaluated/disposition': 'none',\n",
555 " 'feedback/record/row/policy_evaluated/dkim': 'pass',\n",
556 " 'feedback/record/row/policy_evaluated/spf': 'fail',\n",
557 " 'feedback/record/row/source_ip': '65.20.0.12',\n",
558 " 'feedback/report_metadata/date_range/begin': '1459036800',\n",
559 " 'feedback/report_metadata/date_range/end': '1459123199',\n",
560 " 'feedback/report_metadata/email': 'postmaster@dmarc.yahoo.com',\n",
561 " 'feedback/report_metadata/org_name': 'Yahoo! Inc.',\n",
562 " 'feedback/report_metadata/report_id': '1459129809.695034'}"
563 ]
564 },
565 "execution_count": 66,
566 "metadata": {},
567 "output_type": "execute_result"
568 }
569 ],
570 "source": [
571 "xpath_of(root2)"
572 ]
573 },
574 {
575 "cell_type": "code",
576 "execution_count": 63,
577 "metadata": {
578 "collapsed": false
579 },
580 "outputs": [
581 {
582 "data": {
583 "text/plain": [
584 "'1459036800'"
585 ]
586 },
587 "execution_count": 63,
588 "metadata": {},
589 "output_type": "execute_result"
590 }
591 ],
592 "source": [
593 "root2.find('./report_metadata/date_range/begin').text"
594 ]
595 },
596 {
597 "cell_type": "code",
598 "execution_count": 77,
599 "metadata": {
600 "collapsed": false
601 },
602 "outputs": [
603 {
604 "data": {
605 "text/plain": [
606 "'./report_metadata/date_range/begin'"
607 ]
608 },
609 "execution_count": 77,
610 "metadata": {},
611 "output_type": "execute_result"
612 }
613 ],
614 "source": [
615 "re.sub(r'^\\w*(/.*)$', r'.\\1', 'feedback/report_metadata/date_range/begin')"
616 ]
617 },
618 {
619 "cell_type": "code",
620 "execution_count": 82,
621 "metadata": {
622 "collapsed": false
623 },
624 "outputs": [
625 {
626 "data": {
627 "text/plain": [
628 "'report_metadata_date_range_begin'"
629 ]
630 },
631 "execution_count": 82,
632 "metadata": {},
633 "output_type": "execute_result"
634 }
635 ],
636 "source": [
637 "re.sub(r'/', r'_',\n",
638 " re.sub(r'^[^\\/]*/(.*)$', r'\\1', 'feedback/report_metadata/date_range/begin'))"
639 ]
640 },
641 {
642 "cell_type": "code",
643 "execution_count": 156,
644 "metadata": {
645 "collapsed": false,
646 "scrolled": true
647 },
648 "outputs": [
649 {
650 "data": {
651 "text/plain": [
652 "{'./policy_published/adkim': {'pg_field_name': 'policy_published_adkim',\n",
653 " 'pg_table': 'reports',\n",
654 " 'pg_type': 'varchar'},\n",
655 " './policy_published/aspf': {'pg_field_name': 'policy_published_aspf',\n",
656 " 'pg_table': 'reports',\n",
657 " 'pg_type': 'varchar'},\n",
658 " './policy_published/domain': {'pg_field_name': 'policy_published_domain',\n",
659 " 'pg_table': 'reports',\n",
660 " 'pg_type': 'varchar'},\n",
661 " './policy_published/p': {'pg_field_name': 'policy_published_p',\n",
662 " 'pg_table': 'reports',\n",
663 " 'pg_type': 'varchar'},\n",
664 " './policy_published/pct': {'pg_field_name': 'policy_published_pct',\n",
665 " 'pg_table': 'reports',\n",
666 " 'pg_type': 'int'},\n",
667 " './record[{}]/auth_results/dkim/domain': {'pg_field_name': 'auth_results_dkim_domain',\n",
668 " 'pg_table': 'report_items',\n",
669 " 'pg_type': 'varchar'},\n",
670 " './record[{}]/auth_results/dkim/result': {'pg_field_name': 'auth_results_dkim_result',\n",
671 " 'pg_table': 'report_items',\n",
672 " 'pg_type': 'varchar'},\n",
673 " './record[{}]/auth_results/spf/domain': {'pg_field_name': 'auth_results_spf_domain',\n",
674 " 'pg_table': 'report_items',\n",
675 " 'pg_type': 'varchar'},\n",
676 " './record[{}]/auth_results/spf/result': {'pg_field_name': 'auth_results_spf_result',\n",
677 " 'pg_table': 'report_items',\n",
678 " 'pg_type': 'varchar'},\n",
679 " './record[{}]/identifiers/header_from': {'pg_field_name': 'identifiers_header_from',\n",
680 " 'pg_table': 'report_items',\n",
681 " 'pg_type': 'varchar'},\n",
682 " './record[{}]/row/count': {'pg_field_name': 'count',\n",
683 " 'pg_table': 'report_items',\n",
684 " 'pg_type': 'int'},\n",
685 " './record[{}]/row/policy_evaluated/disposition': {'pg_field_name': 'policy_evaluated_disposition',\n",
686 " 'pg_table': 'report_items',\n",
687 " 'pg_type': 'varchar'},\n",
688 " './record[{}]/row/policy_evaluated/dkim': {'pg_field_name': 'policy_evaluated_dkim',\n",
689 " 'pg_table': 'report_items',\n",
690 " 'pg_type': 'varchar'},\n",
691 " './record[{}]/row/policy_evaluated/spf': {'pg_field_name': 'policy_evaluated_spf',\n",
692 " 'pg_table': 'report_items',\n",
693 " 'pg_type': 'varchar'},\n",
694 " './record[{}]/row/source_ip': {'pg_field_name': 'source_ip',\n",
695 " 'pg_table': 'report_items',\n",
696 " 'pg_type': 'inet'},\n",
697 " './report_metadata/date_range/begin': {'pg_field_name': 'report_metadata_date_range_begin',\n",
698 " 'pg_table': 'reports',\n",
699 " 'pg_type': 'timestamp'},\n",
700 " './report_metadata/date_range/end': {'pg_field_name': 'report_metadata_date_range_end',\n",
701 " 'pg_table': 'reports',\n",
702 " 'pg_type': 'timestamp'},\n",
703 " './report_metadata/email': {'pg_field_name': 'report_metadata_email',\n",
704 " 'pg_table': 'reports',\n",
705 " 'pg_type': 'varchar'},\n",
706 " './report_metadata/org_name': {'pg_field_name': 'report_metadata_org_name',\n",
707 " 'pg_table': 'reports',\n",
708 " 'pg_type': 'varchar'},\n",
709 " './report_metadata/report_id': {'pg_field_name': 'report_metadata_report_id',\n",
710 " 'pg_table': 'reports',\n",
711 " 'pg_type': 'varchar'}}"
712 ]
713 },
714 "execution_count": 156,
715 "metadata": {},
716 "output_type": "execute_result"
717 }
718 ],
719 "source": [
720 "field_maps = {re.sub(r'^\\w*(/.*)$', r'.\\1', name): \n",
721 " {'pg_field_name': re.sub(r'/', r'_', re.sub(r'^[^\\/]*/(.*)$', r'\\1', name)),\n",
722 " 'pg_table': 'reports',\n",
723 " 'pg_type': 'varchar'}\n",
724 " for name in xpath_of(root2)\n",
725 " if 'record' not in name}\n",
726 "field_maps.update({re.sub(r'^\\w*/record(/.*)$', r'./record[{}]\\1', name): \n",
727 " {'pg_field_name': re.sub(r'/', r'_', re.sub(r'^.*/record(/row)?/(.*)$', r'\\2', name)),\n",
728 " 'pg_table': 'report_items',\n",
729 " 'pg_type': 'varchar'}\n",
730 " for name in xpath_of(root2)\n",
731 " if 'record' in name})\n",
732 "field_maps['./report_metadata/date_range/begin']['pg_type'] = 'timestamp'\n",
733 "field_maps['./report_metadata/date_range/end']['pg_type'] = 'timestamp'\n",
734 "field_maps['./policy_published/pct']['pg_type'] = 'int'\n",
735 "field_maps['./record[{}]/row/count']['pg_type'] = 'int'\n",
736 "field_maps['./record[{}]/row/source_ip']['pg_type'] = 'inet'\n",
737 "\n",
738 "field_maps"
739 ]
740 },
741 {
742 "cell_type": "code",
743 "execution_count": 95,
744 "metadata": {
745 "collapsed": false
746 },
747 "outputs": [
748 {
749 "data": {
750 "text/plain": [
751 "datetime.datetime(2016, 3, 27, 23, 59, 59)"
752 ]
753 },
754 "execution_count": 95,
755 "metadata": {},
756 "output_type": "execute_result"
757 }
758 ],
759 "source": [
760 "datetime.datetime.utcfromtimestamp(1459123199)"
761 ]
762 },
763 {
764 "cell_type": "code",
765 "execution_count": 100,
766 "metadata": {
767 "collapsed": false
768 },
769 "outputs": [
770 {
771 "data": {
772 "text/plain": [
773 "1459119599.0"
774 ]
775 },
776 "execution_count": 100,
777 "metadata": {},
778 "output_type": "execute_result"
779 }
780 ],
781 "source": [
782 "datetime.datetime.utcfromtimestamp(1459123199).timestamp()"
783 ]
784 },
785 {
786 "cell_type": "code",
787 "execution_count": 157,
788 "metadata": {
789 "collapsed": false
790 },
791 "outputs": [
792 {
793 "data": {
794 "text/plain": [
795 "'create table reports (id serial primary key, policy_published_p varchar, report_metadata_date_range_end timestamp, report_metadata_org_name varchar, policy_published_aspf varchar, policy_published_adkim varchar, policy_published_pct int, report_metadata_email varchar, report_metadata_date_range_begin timestamp, report_metadata_report_id varchar, policy_published_domain varchar);'"
796 ]
797 },
798 "execution_count": 157,
799 "metadata": {},
800 "output_type": "execute_result"
801 }
802 ],
803 "source": [
804 "create_report_table_string = 'create table reports (id serial primary key, ' + \\\n",
805 "', '.join(field_maps[p]['pg_field_name'] + ' ' + field_maps[p]['pg_type'] \n",
806 " for p in field_maps if field_maps[p]['pg_table'] == 'reports') + \\\n",
807 "');'\n",
808 "create_report_table_string"
809 ]
810 },
811 {
812 "cell_type": "code",
813 "execution_count": 193,
814 "metadata": {
815 "collapsed": false
816 },
817 "outputs": [
818 {
819 "data": {
820 "text/plain": [
821 "'create table report_items (id serial primary key, report_id integer references reports, auth_results_spf_domain varchar, auth_results_dkim_domain varchar, policy_evaluated_dkim varchar, auth_results_spf_result varchar, source_ip inet, policy_evaluated_spf varchar, count int, identifiers_header_from varchar, policy_evaluated_disposition varchar, auth_results_dkim_result varchar);'"
822 ]
823 },
824 "execution_count": 193,
825 "metadata": {},
826 "output_type": "execute_result"
827 }
828 ],
829 "source": [
830 "create_report_item_table_string = 'create table report_items (id serial primary key, ' + \\\n",
831 " 'report_id integer references reports, ' + \\\n",
832 "', '.join(field_maps[p]['pg_field_name'] + ' ' + field_maps[p]['pg_type'] \n",
833 " for p in field_maps if field_maps[p]['pg_table'] == 'report_items') + \\\n",
834 "');'\n",
835 "create_report_item_table_string"
836 ]
837 },
838 {
839 "cell_type": "code",
840 "execution_count": 117,
841 "metadata": {
842 "collapsed": false
843 },
844 "outputs": [
845 {
846 "data": {
847 "text/plain": [
848 "<cursor object at 0x7fa9a819b7c8; closed: 0>"
849 ]
850 },
851 "execution_count": 117,
852 "metadata": {},
853 "output_type": "execute_result"
854 }
855 ],
856 "source": [
857 "cur"
858 ]
859 },
860 {
861 "cell_type": "code",
862 "execution_count": 194,
863 "metadata": {
864 "collapsed": false
865 },
866 "outputs": [],
867 "source": [
868 "cur.execute(create_report_table_string)"
869 ]
870 },
871 {
872 "cell_type": "code",
873 "execution_count": 195,
874 "metadata": {
875 "collapsed": false
876 },
877 "outputs": [],
878 "source": [
879 "cur.execute(create_report_item_table_string)"
880 ]
881 },
882 {
883 "cell_type": "code",
884 "execution_count": 196,
885 "metadata": {
886 "collapsed": true
887 },
888 "outputs": [],
889 "source": [
890 "conn.commit()"
891 ]
892 },
893 {
894 "cell_type": "code",
895 "execution_count": 128,
896 "metadata": {
897 "collapsed": false
898 },
899 "outputs": [],
900 "source": [
901 "rep = dmarc_reports['google.com!njae.me.uk!1458777600!1458863999.xml']"
902 ]
903 },
904 {
905 "cell_type": "code",
906 "execution_count": 134,
907 "metadata": {
908 "collapsed": false
909 },
910 "outputs": [
911 {
912 "data": {
913 "text/plain": [
914 "'1'"
915 ]
916 },
917 "execution_count": 134,
918 "metadata": {},
919 "output_type": "execute_result"
920 }
921 ],
922 "source": [
923 "rep.find('./record/row/count').text"
924 ]
925 },
926 {
927 "cell_type": "code",
928 "execution_count": 135,
929 "metadata": {
930 "collapsed": false
931 },
932 "outputs": [
933 {
934 "data": {
935 "text/plain": [
936 "[('google.com!njae.me.uk!1458950400!1459036799.xml', '1'),\n",
937 " ('yahoo.com!njae.me.uk!1459036800!1459123199.xml', '1'),\n",
938 " ('yahoo.com!njae.me.uk!1458864000!1458950399.xml', '2'),\n",
939 " ('hotmail.com!njae.me.uk!1459011600!1459098000.xml', '2'),\n",
940 " ('google.com!njae.me.uk!1458777600!1458863999.xml', '1')]"
941 ]
942 },
943 "execution_count": 135,
944 "metadata": {},
945 "output_type": "execute_result"
946 }
947 ],
948 "source": [
949 "[(r, dmarc_reports[r].find('./record/row/count').text) for r in dmarc_reports]"
950 ]
951 },
952 {
953 "cell_type": "code",
954 "execution_count": 136,
955 "metadata": {
956 "collapsed": false
957 },
958 "outputs": [
959 {
960 "name": "stdout",
961 "output_type": "stream",
962 "text": [
963 " report_metadata : \n",
964 " org_name : Microsoft Corp.\n",
965 " email : dmarcrep@microsoft.com\n",
966 " report_id : 68aad5080a774e2c997d159b546569b9@hotmail.com\n",
967 " date_range : \n",
968 " begin : 1459011600\n",
969 " end : 1459098000\n",
970 " policy_published : \n",
971 " domain : njae.me.uk\n",
972 " adkim : r\n",
973 " aspf : r\n",
974 " p : none\n",
975 " sp : none\n",
976 " pct : 100\n",
977 " record : \n",
978 " row : \n",
979 " source_ip : 212.69.55.62\n",
980 " count : 2\n",
981 " policy_evaluated : \n",
982 " disposition : none\n",
983 " dkim : pass\n",
984 " spf : pass\n",
985 " identifiers : \n",
986 " header_from : njae.me.uk\n",
987 " auth_results : \n",
988 " spf : \n",
989 " domain : njae.me.uk\n",
990 " result : pass\n",
991 " dkim : \n",
992 " domain : njae.me.uk\n",
993 " result : pass\n"
994 ]
995 }
996 ],
997 "source": [
998 "walk(dmarc_reports['hotmail.com!njae.me.uk!1459011600!1459098000.xml'])"
999 ]
1000 },
1001 {
1002 "cell_type": "code",
1003 "execution_count": 137,
1004 "metadata": {
1005 "collapsed": false
1006 },
1007 "outputs": [
1008 {
1009 "name": "stdout",
1010 "output_type": "stream",
1011 "text": [
1012 " report_metadata : \n",
1013 " org_name : google.com\n",
1014 " email : noreply-dmarc-support@google.com\n",
1015 " extra_contact_info : https://support.google.com/a/answer/2466580\n",
1016 " report_id : 2150510829392606201\n",
1017 " date_range : \n",
1018 " begin : 1458950400\n",
1019 " end : 1459036799\n",
1020 " policy_published : \n",
1021 " domain : njae.me.uk\n",
1022 " adkim : r\n",
1023 " aspf : r\n",
1024 " p : none\n",
1025 " sp : none\n",
1026 " pct : 100\n",
1027 " record : \n",
1028 " row : \n",
1029 " source_ip : 82.109.184.9\n",
1030 " count : 1\n",
1031 " policy_evaluated : \n",
1032 " disposition : none\n",
1033 " dkim : fail\n",
1034 " spf : fail\n",
1035 " identifiers : \n",
1036 " header_from : njae.me.uk\n",
1037 " auth_results : \n",
1038 " dkim : \n",
1039 " domain : clublloyds.com\n",
1040 " result : pass\n",
1041 " spf : \n",
1042 " domain : clublloyds.com\n",
1043 " result : pass\n",
1044 " record : \n",
1045 " row : \n",
1046 " source_ip : 212.69.55.62\n",
1047 " count : 1\n",
1048 " policy_evaluated : \n",
1049 " disposition : none\n",
1050 " dkim : pass\n",
1051 " spf : pass\n",
1052 " identifiers : \n",
1053 " header_from : njae.me.uk\n",
1054 " auth_results : \n",
1055 " dkim : \n",
1056 " domain : njae.me.uk\n",
1057 " result : pass\n",
1058 " spf : \n",
1059 " domain : njae.me.uk\n",
1060 " result : pass\n"
1061 ]
1062 }
1063 ],
1064 "source": [
1065 "walk(dmarc_reports['google.com!njae.me.uk!1458950400!1459036799.xml'])"
1066 ]
1067 },
1068 {
1069 "cell_type": "code",
1070 "execution_count": 159,
1071 "metadata": {
1072 "collapsed": false
1073 },
1074 "outputs": [
1075 {
1076 "data": {
1077 "text/plain": [
1078 "['./record[{}]/auth_results/spf/domain',\n",
1079 " './record[{}]/auth_results/dkim/domain',\n",
1080 " './record[{}]/row/policy_evaluated/dkim',\n",
1081 " './record[{}]/auth_results/spf/result',\n",
1082 " './record[{}]/row/source_ip',\n",
1083 " './record[{}]/row/policy_evaluated/spf',\n",
1084 " './record[{}]/row/count',\n",
1085 " './record[{}]/identifiers/header_from',\n",
1086 " './record[{}]/row/policy_evaluated/disposition',\n",
1087 " './record[{}]/auth_results/dkim/result']"
1088 ]
1089 },
1090 "execution_count": 159,
1091 "metadata": {},
1092 "output_type": "execute_result"
1093 }
1094 ],
1095 "source": [
1096 "[p for p in field_maps if field_maps[p]['pg_table'] == 'report_items']"
1097 ]
1098 },
1099 {
1100 "cell_type": "code",
1101 "execution_count": 140,
1102 "metadata": {
1103 "collapsed": false
1104 },
1105 "outputs": [
1106 {
1107 "data": {
1108 "text/plain": [
1109 "2"
1110 ]
1111 },
1112 "execution_count": 140,
1113 "metadata": {},
1114 "output_type": "execute_result"
1115 }
1116 ],
1117 "source": [
1118 "len(dmarc_reports['google.com!njae.me.uk!1458950400!1459036799.xml'].findall('./record'))"
1119 ]
1120 },
1121 {
1122 "cell_type": "code",
1123 "execution_count": 144,
1124 "metadata": {
1125 "collapsed": false
1126 },
1127 "outputs": [
1128 {
1129 "data": {
1130 "text/plain": [
1131 "'212.69.55.62'"
1132 ]
1133 },
1134 "execution_count": 144,
1135 "metadata": {},
1136 "output_type": "execute_result"
1137 }
1138 ],
1139 "source": [
1140 "dmarc_reports['google.com!njae.me.uk!1458950400!1459036799.xml'].find('./record[2]/row/source_ip').text"
1141 ]
1142 },
1143 {
1144 "cell_type": "code",
1145 "execution_count": 153,
1146 "metadata": {
1147 "collapsed": false
1148 },
1149 "outputs": [
1150 {
1151 "data": {
1152 "text/plain": [
1153 "'82.109.184.9'"
1154 ]
1155 },
1156 "execution_count": 153,
1157 "metadata": {},
1158 "output_type": "execute_result"
1159 }
1160 ],
1161 "source": [
1162 "dmarc_reports['google.com!njae.me.uk!1458950400!1459036799.xml'].find('./record[{}]/row/source_ip'.format(1)).text"
1163 ]
1164 },
1165 {
1166 "cell_type": "code",
1167 "execution_count": 163,
1168 "metadata": {
1169 "collapsed": false
1170 },
1171 "outputs": [
1172 {
1173 "name": "stdout",
1174 "output_type": "stream",
1175 "text": [
1176 "policy_published_p :> none\n",
1177 "report_metadata_date_range_end :> 1459036799\n",
1178 "report_metadata_org_name :> google.com\n",
1179 "policy_published_aspf :> r\n",
1180 "policy_published_adkim :> r\n",
1181 "policy_published_pct :> 100\n",
1182 "report_metadata_email :> noreply-dmarc-support@google.com\n",
1183 "report_metadata_date_range_begin :> 1458950400\n",
1184 "report_metadata_report_id :> 2150510829392606201\n",
1185 "policy_published_domain :> njae.me.uk\n"
1186 ]
1187 }
1188 ],
1189 "source": [
1190 "rep = dmarc_reports['google.com!njae.me.uk!1458950400!1459036799.xml']\n",
1191 "for f in [f for f in field_maps if field_maps[f]['pg_table'] == 'reports']:\n",
1192 " print(field_maps[f]['pg_field_name'], ':>', rep.find(f).text)"
1193 ]
1194 },
1195 {
1196 "cell_type": "code",
1197 "execution_count": 165,
1198 "metadata": {
1199 "collapsed": false
1200 },
1201 "outputs": [
1202 {
1203 "name": "stdout",
1204 "output_type": "stream",
1205 "text": [
1206 "1\n",
1207 "auth_results_spf_domain :> clublloyds.com\n",
1208 "auth_results_dkim_domain :> clublloyds.com\n",
1209 "policy_evaluated_dkim :> fail\n",
1210 "auth_results_spf_result :> pass\n",
1211 "source_ip :> 82.109.184.9\n",
1212 "policy_evaluated_spf :> fail\n",
1213 "count :> 1\n",
1214 "identifiers_header_from :> njae.me.uk\n",
1215 "policy_evaluated_disposition :> none\n",
1216 "auth_results_dkim_result :> pass\n",
1217 "2\n",
1218 "auth_results_spf_domain :> njae.me.uk\n",
1219 "auth_results_dkim_domain :> njae.me.uk\n",
1220 "policy_evaluated_dkim :> pass\n",
1221 "auth_results_spf_result :> pass\n",
1222 "source_ip :> 212.69.55.62\n",
1223 "policy_evaluated_spf :> pass\n",
1224 "count :> 1\n",
1225 "identifiers_header_from :> njae.me.uk\n",
1226 "policy_evaluated_disposition :> none\n",
1227 "auth_results_dkim_result :> pass\n"
1228 ]
1229 }
1230 ],
1231 "source": [
1232 "rep = dmarc_reports['google.com!njae.me.uk!1458950400!1459036799.xml']\n",
1233 "for i in range(1, len(rep.findall('./record'))+1):\n",
1234 " print(i)\n",
1235 " for f in [f for f in field_maps if field_maps[f]['pg_table'] == 'report_items']:\n",
1236 " print(field_maps[f]['pg_field_name'], ':>', rep.find(f.format(i)).text)"
1237 ]
1238 },
1239 {
1240 "cell_type": "code",
1241 "execution_count": 166,
1242 "metadata": {
1243 "collapsed": true
1244 },
1245 "outputs": [],
1246 "source": [
1247 "def maybe_strip(text):\n",
1248 " if text:\n",
1249 " return text.strip()\n",
1250 " else:\n",
1251 " return ''"
1252 ]
1253 },
1254 {
1255 "cell_type": "code",
1256 "execution_count": 197,
1257 "metadata": {
1258 "collapsed": false
1259 },
1260 "outputs": [
1261 {
1262 "data": {
1263 "text/plain": [
1264 "('insert into reports (policy_published_p, report_metadata_date_range_end, report_metadata_org_name, policy_published_aspf, policy_published_adkim, policy_published_pct, report_metadata_email, report_metadata_date_range_begin, report_metadata_report_id, policy_published_domain) values (%(policy_published_p)s, %(report_metadata_date_range_end)s, %(report_metadata_org_name)s, %(policy_published_aspf)s, %(policy_published_adkim)s, %(policy_published_pct)s, %(report_metadata_email)s, %(report_metadata_date_range_begin)s, %(report_metadata_report_id)s, %(policy_published_domain)s);',\n",
1265 " {'policy_published_adkim': 'r',\n",
1266 " 'policy_published_aspf': 'r',\n",
1267 " 'policy_published_domain': 'njae.me.uk',\n",
1268 " 'policy_published_p': 'none',\n",
1269 " 'policy_published_pct': 100,\n",
1270 " 'report_metadata_date_range_begin': datetime.datetime(2016, 3, 26, 0, 0),\n",
1271 " 'report_metadata_date_range_end': datetime.datetime(2016, 3, 26, 23, 59, 59),\n",
1272 " 'report_metadata_email': 'noreply-dmarc-support@google.com',\n",
1273 " 'report_metadata_org_name': 'google.com',\n",
1274 " 'report_metadata_report_id': '2150510829392606201'})"
1275 ]
1276 },
1277 "execution_count": 197,
1278 "metadata": {},
1279 "output_type": "execute_result"
1280 }
1281 ],
1282 "source": [
1283 "rep = dmarc_reports['google.com!njae.me.uk!1458950400!1459036799.xml']\n",
1284 "field_names = []\n",
1285 "values = {}\n",
1286 "for f in [f for f in field_maps if field_maps[f]['pg_table'] == 'reports']:\n",
1287 " field_names += [field_maps[f]['pg_field_name']]\n",
1288 " if field_maps[f]['pg_type'] == 'int':\n",
1289 " values[field_maps[f]['pg_field_name']] = int(rep.find(f).text)\n",
1290 " elif field_maps[f]['pg_type'] == 'timestamp':\n",
1291 " values[field_maps[f]['pg_field_name']] = datetime.datetime.utcfromtimestamp(int(rep.find(f).text))\n",
1292 " elif field_maps[f]['pg_type'] == 'inet':\n",
1293 " values[field_maps[f]['pg_field_name']] = maybe_strip(rep.find(f).text)\n",
1294 " else:\n",
1295 " values[field_maps[f]['pg_field_name']] = maybe_strip(rep.find(f).text)\n",
1296 "insert_string = 'insert into reports (' + ', '.join(field_names) +') '\n",
1297 "insert_string += 'values (' + ', '.join('%({})s'.format(f) for f in field_names) + ');'\n",
1298 "insert_string, values"
1299 ]
1300 },
1301 {
1302 "cell_type": "code",
1303 "execution_count": 198,
1304 "metadata": {
1305 "collapsed": false
1306 },
1307 "outputs": [],
1308 "source": [
1309 "cur.execute(insert_string, values)\n",
1310 "conn.commit()"
1311 ]
1312 },
1313 {
1314 "cell_type": "code",
1315 "execution_count": 201,
1316 "metadata": {
1317 "collapsed": false
1318 },
1319 "outputs": [
1320 {
1321 "data": {
1322 "text/plain": [
1323 "1"
1324 ]
1325 },
1326 "execution_count": 201,
1327 "metadata": {},
1328 "output_type": "execute_result"
1329 }
1330 ],
1331 "source": [
1332 "cur.execute('select id, report_metadata_report_id from reports where report_metadata_report_id = %s;', \n",
1333 " [rep.find('./report_metadata/report_id').text])\n",
1334 "cur.fetchall()[0][0]"
1335 ]
1336 },
1337 {
1338 "cell_type": "code",
1339 "execution_count": 203,
1340 "metadata": {
1341 "collapsed": false
1342 },
1343 "outputs": [
1344 {
1345 "name": "stdout",
1346 "output_type": "stream",
1347 "text": [
1348 "1\n",
1349 "insert into report_items (report_id, auth_results_spf_domain, auth_results_dkim_domain, policy_evaluated_dkim, auth_results_spf_result, source_ip, policy_evaluated_spf, count, identifiers_header_from, policy_evaluated_disposition, auth_results_dkim_result) values (%(report_id)s, %(auth_results_spf_domain)s, %(auth_results_dkim_domain)s, %(policy_evaluated_dkim)s, %(auth_results_spf_result)s, %(source_ip)s, %(policy_evaluated_spf)s, %(count)s, %(identifiers_header_from)s, %(policy_evaluated_disposition)s, %(auth_results_dkim_result)s); {'auth_results_spf_result': 'pass', 'count': 1, 'report_id': 1, 'identifiers_header_from': 'njae.me.uk', 'policy_evaluated_disposition': 'none', 'auth_results_dkim_result': 'pass', 'source_ip': '82.109.184.9', 'policy_evaluated_spf': 'fail', 'auth_results_dkim_domain': 'clublloyds.com', 'auth_results_spf_domain': 'clublloyds.com', 'policy_evaluated_dkim': 'fail'}\n",
1350 "2\n",
1351 "insert into report_items (report_id, auth_results_spf_domain, auth_results_dkim_domain, policy_evaluated_dkim, auth_results_spf_result, source_ip, policy_evaluated_spf, count, identifiers_header_from, policy_evaluated_disposition, auth_results_dkim_result) values (%(report_id)s, %(auth_results_spf_domain)s, %(auth_results_dkim_domain)s, %(policy_evaluated_dkim)s, %(auth_results_spf_result)s, %(source_ip)s, %(policy_evaluated_spf)s, %(count)s, %(identifiers_header_from)s, %(policy_evaluated_disposition)s, %(auth_results_dkim_result)s); {'auth_results_spf_result': 'pass', 'count': 1, 'report_id': 1, 'identifiers_header_from': 'njae.me.uk', 'policy_evaluated_disposition': 'none', 'auth_results_dkim_result': 'pass', 'source_ip': '212.69.55.62', 'policy_evaluated_spf': 'pass', 'auth_results_dkim_domain': 'njae.me.uk', 'auth_results_spf_domain': 'njae.me.uk', 'policy_evaluated_dkim': 'pass'}\n"
1352 ]
1353 }
1354 ],
1355 "source": [
1356 "rep = dmarc_reports['google.com!njae.me.uk!1458950400!1459036799.xml']\n",
1357 "for i in range(1, len(rep.findall('./record'))+1):\n",
1358 " print(i)\n",
1359 " field_names = []\n",
1360 " cur.execute('select id, report_metadata_report_id from reports where report_metadata_report_id = %s;', \n",
1361 " [rep.find('./report_metadata/report_id').text])\n",
1362 " results = cur.fetchall()\n",
1363 " if len(results) != 1:\n",
1364 " raise\n",
1365 " else:\n",
1366 " report_id = results[0][0]\n",
1367 " values = {'report_id': report_id}\n",
1368 " for f in [f for f in field_maps if field_maps[f]['pg_table'] == 'report_items']:\n",
1369 " field_names += [field_maps[f]['pg_field_name']]\n",
1370 " if field_maps[f]['pg_type'] == 'int':\n",
1371 " values[field_maps[f]['pg_field_name']] = int(rep.find(f.format(i)).text)\n",
1372 " elif field_maps[f]['pg_type'] == 'timestamp':\n",
1373 " values[field_maps[f]['pg_field_name']] = datetime.datetime.utcfromtimestamp(int(rep.find(f.format(i)).text))\n",
1374 " elif field_maps[f]['pg_type'] == 'inet':\n",
1375 " values[field_maps[f]['pg_field_name']] = maybe_strip(rep.find(f.format(i)).text)\n",
1376 " else:\n",
1377 " values[field_maps[f]['pg_field_name']] = maybe_strip(rep.find(f.format(i)).text)\n",
1378 " insert_string = 'insert into report_items (report_id, ' + ', '.join(field_names) +') '\n",
1379 " insert_string += 'values (%(report_id)s, ' + ', '.join('%({})s'.format(f) for f in field_names) + ');'\n",
1380 " print(insert_string, values)\n",
1381 " cur.execute(insert_string, values)\n",
1382 "conn.commit()"
1383 ]
1384 },
1385 {
1386 "cell_type": "code",
1387 "execution_count": 191,
1388 "metadata": {
1389 "collapsed": true
1390 },
1391 "outputs": [],
1392 "source": [
1393 "conn.rollback()"
1394 ]
1395 },
1396 {
1397 "cell_type": "code",
1398 "execution_count": 209,
1399 "metadata": {
1400 "collapsed": true
1401 },
1402 "outputs": [],
1403 "source": [
1404 "def write_report(connection, cursor, report):\n",
1405 " \n",
1406 " field_names = []\n",
1407 " values = {}\n",
1408 " for f in [f for f in field_maps if field_maps[f]['pg_table'] == 'reports']:\n",
1409 " field_names += [field_maps[f]['pg_field_name']]\n",
1410 " if field_maps[f]['pg_type'] == 'int':\n",
1411 " values[field_maps[f]['pg_field_name']] = int(report.find(f).text)\n",
1412 " elif field_maps[f]['pg_type'] == 'timestamp':\n",
1413 " values[field_maps[f]['pg_field_name']] = datetime.datetime.utcfromtimestamp(int(report.find(f).text))\n",
1414 " elif field_maps[f]['pg_type'] == 'inet':\n",
1415 " values[field_maps[f]['pg_field_name']] = maybe_strip(report.find(f).text)\n",
1416 " else:\n",
1417 " values[field_maps[f]['pg_field_name']] = maybe_strip(report.find(f).text)\n",
1418 " insert_string = 'insert into reports (' + ', '.join(field_names) + ') '\n",
1419 " insert_string += 'values (' + ', '.join('%({})s'.format(f) for f in field_names) + ');'\n",
1420 " cursor.execute(insert_string, values)\n",
1421 " \n",
1422 " for i in range(1, len(report.findall('./record'))+1):\n",
1423 " field_names = []\n",
1424 " cursor.execute('select id, report_metadata_report_id from reports where report_metadata_report_id = %s;', \n",
1425 " [report.find('./report_metadata/report_id').text])\n",
1426 " results = cursor.fetchall()\n",
1427 " if len(results) != 1:\n",
1428 " raise\n",
1429 " else:\n",
1430 " report_id = results[0][0]\n",
1431 " values = {'report_id': report_id}\n",
1432 " for f in [f for f in field_maps if field_maps[f]['pg_table'] == 'report_items']:\n",
1433 " field_names += [field_maps[f]['pg_field_name']]\n",
1434 " if field_maps[f]['pg_type'] == 'int':\n",
1435 " values[field_maps[f]['pg_field_name']] = int(report.find(f.format(i)).text)\n",
1436 " elif field_maps[f]['pg_type'] == 'timestamp':\n",
1437 " values[field_maps[f]['pg_field_name']] = datetime.datetime.utcfromtimestamp(int(report.find(f.format(i)).text))\n",
1438 " elif field_maps[f]['pg_type'] == 'inet':\n",
1439 " values[field_maps[f]['pg_field_name']] = maybe_strip(report.find(f.format(i)).text)\n",
1440 " else:\n",
1441 " values[field_maps[f]['pg_field_name']] = maybe_strip(report.find(f.format(i)).text)\n",
1442 " insert_string = 'insert into report_items (report_id, ' + ', '.join(field_names) + ') '\n",
1443 " insert_string += 'values (%(report_id)s, ' + ', '.join('%({})s'.format(f) for f in field_names) + ');'\n",
1444 " cursor.execute(insert_string, values)\n",
1445 " connection.commit()"
1446 ]
1447 },
1448 {
1449 "cell_type": "code",
1450 "execution_count": 210,
1451 "metadata": {
1452 "collapsed": false
1453 },
1454 "outputs": [],
1455 "source": [
1456 "for rep in dmarc_reports:\n",
1457 " write_report(conn, cur, dmarc_reports[rep])"
1458 ]
1459 },
1460 {
1461 "cell_type": "code",
1462 "execution_count": 211,
1463 "metadata": {
1464 "collapsed": true
1465 },
1466 "outputs": [],
1467 "source": [
1468 "conn.close()"
1469 ]
1470 },
1471 {
1472 "cell_type": "code",
1473 "execution_count": 208,
1474 "metadata": {
1475 "collapsed": true
1476 },
1477 "outputs": [],
1478 "source": [
1479 "conn.rollback()"
1480 ]
1481 },
1482 {
1483 "cell_type": "code",
1484 "execution_count": null,
1485 "metadata": {
1486 "collapsed": true
1487 },
1488 "outputs": [],
1489 "source": []
1490 }
1491 ],
1492 "metadata": {
1493 "kernelspec": {
1494 "display_name": "Python 3",
1495 "language": "python",
1496 "name": "python3"
1497 },
1498 "language_info": {
1499 "codemirror_mode": {
1500 "name": "ipython",
1501 "version": 3
1502 },
1503 "file_extension": ".py",
1504 "mimetype": "text/x-python",
1505 "name": "python",
1506 "nbconvert_exporter": "python",
1507 "pygments_lexer": "ipython3",
1508 "version": "3.4.3+"
1509 }
1510 },
1511 "nbformat": 4,
1512 "nbformat_minor": 0
1513 }