Permalink
Find file
Fetching contributors…
Cannot retrieve contributors at this time
155 lines (154 sloc) 7.02 KB
[
{
"api_url": "http://your-server",
"base_url": "http://your-server",
"es_paper_index": "paper",
"file_extensions": [
[
"application/pdf",
"pdf"
],
[
"image/tiff",
"tif"
],
[
"image/jpeg",
"jpg"
],
[
"application/vnd.ms-powerpoint",
"pptx"
],
[
"application/msword",
"doc"
],
[
"application/zip",
"zip"
]
],
"file_folder": "/srv/www/ris-web/webapp/static/files/",
"gs_cmd": "/usr/bin/gs",
"map_attribution": "Map Data © <a href='http://www.openstreetmap.org'>OpenStreetMap</a> contributors, Tiles courtesy of <a href='http://www.mapquest.com/' target='_blank'>MapQuest</a>.",
"map_tile_url_schema": "http://otile1.mqcdn.com/tiles/1.0.0/map/{z}/{x}/{y}.jpg",
"map_zoomlevel_max": 18,
"map_zoomlevel_min": "4",
"pdf_to_text_cmd": "/usr/bin/pdftotext",
"region_default": "5482ec449bcda40691218ef1",
"scraper": {
"log_base_dir": ".",
"log_level": "DEBUG",
"sessionnet-asp": {
"urls": {
"AGENDAITEM_DETAIL_PARSE_PATTERN": "to0050.asp?__ktonr={agendaitem_id:d}",
"AGENDAITEM_DETAIL_PRINT_PATTERN": "%sto0050.asp?__ktonr=%d",
"CALENDAR_MONTH_PARSE_PATTERN": "si0040.asp?__cjahr={year:d}&__cmonat={month:d}",
"CALENDAR_MONTH_PRINT_PATTERN": "%ssi0040.asp?__cjahr=%d&__cmonat=%d",
"FILE_DOWNLOAD_TARGET": [
"ydocstart.asp",
"getfile.asp"
],
"MEETING_DETAIL_PARSE_PATTERN": "to0040.asp?__ksinr={meeting_id:d}",
"MEETING_DETAIL_PRINT_PATTERN": "%sto0040.asp?__ksinr=%d",
"ORGANIZATION_DETAIL_PARSE_PATTERN": "kp0040.asp?__kgrnr={committee_id:d}",
"ORGANIZATION_DETAIL_PARSE_PATTERN_FULL": "kp0040.asp?__cwp=1&__kgrnr={committee_id:d}",
"ORGANIZATION_DETAIL_PRINT_PATTERN": "%skp0040.asp?__kgrnr=%d",
"ORGANIZATION_DETAIL_PRINT_PATTERN_FULL": "%skp0040.asp?__cwpall=1&__kgrnr=%d",
"PAPER_DETAIL_PARSE_PATTERN": "vo0050.asp?__kvonr={paper_id:d}",
"PAPER_DETAIL_PRINT_PATTERN": "%svo0050.asp?__kvonr=%d",
"PERSON_DETAIL_PARSE_PATTERN": "kp0051.asp?__kpenr={person_id:d}",
"PERSON_DETAIL_PARSE_PATTERN_ALT": "kp0050.asp?__kpenr={person_id:d}&grnr=0",
"PERSON_DETAIL_PRINT_PATTERN": "%skp0051.asp?__kpenr=%d",
"PERSON_DETAIL_PRINT_PATTERN_ALT": "%skp0050.asp?__kpenr=%d&grnr=0",
"PERSON_ORGANIZATION_PARSE_PATTERN": "kp0050.asp?__cwpall=1&__kpenr={person_id:d}",
"PERSON_ORGANIZATION_PRINT_PATTERN": "%skp0050.asp?__cwpall=1&__kpenr=%d",
"PERSON_OVERVIEW_PARSE_PATTERN": "kp0041.asp?__cwpall=1",
"PERSON_OVERVIEW_PRINT_PATTERN": "%skp0041.asp?__cwpall=1"
},
"xpath": {
"MEETING_DETAIL_AGENDAITEM_ROWS": "//*[@class='smccontenttable smc_page_to0040_contenttable']/tbody/tr",
"MEETING_DETAIL_AGENDAITEM_ROWS_PAPER_LINK": ".//a",
"MEETING_DETAIL_FILES": "//*[@id='smccontent']//table",
"MEETING_DETAIL_FILES_CONTAINER_CLASSNAME": "smcdocbox",
"MEETING_DETAIL_IDENTIFIER_TD": "//*[@id=\"smctablevorgang\"]/tbody//td",
"MEETING_DETAIL_ORGANIZATION_LINK": "//li[@class='smcmenucontext_fct_gremium']/a",
"MEETING_DETAIL_TITLE": "//h1",
"PAPER_DETAIL_AGENDA_ROWS": "//*[@class='smccontenttable smc_page_vo0050_contenttable']/tbody/tr",
"PAPER_DETAIL_FILES": "//*[@id='smccontent']//table",
"PAPER_DETAIL_FILES_CONTAINER_CLASSNAME": "smcdocbox",
"PAPER_DETAIL_IDENTIFIER_TD": "//*[@id='smctablevorgang']/tbody//td",
"PAPER_DETAIL_TITLE": "//h1",
"PERSONLIST_LINES": "//table[@id='smc_page_kp0041_contenttable1']//tr",
"PERSON_ORGANIZATION_LINES": "//table[@id='smc_page_kp0050_contenttable1']//tr"
}
},
"sessionnet-php": {
"urls": {
"AGENDAITEM_DETAIL_PARSE_PATTERN": "to0050.php?__ktonr={agendaitem_id:d}",
"AGENDAITEM_DETAIL_PRINT_PATTERN": "%sto0050.php?__ktonr=%d",
"CALENDAR_MONTH_PARSE_PATTERN": "si0040.php?__cjahr={year:d}&__cmonat={month:d}",
"CALENDAR_MONTH_PRINT_PATTERN": "%ssi0040.php?__cjahr=%d&__cmonat=%d",
"FILE_DOWNLOAD_TARGET": [
"ydocstart.php",
"getfile.php"
],
"MEETING_DETAIL_PARSE_PATTERN": "to0040.php?__ksinr={meeting_id:d}",
"MEETING_DETAIL_PRINT_PATTERN": "%sto0040.php?__ksinr=%d",
"ORGANIZATION_DETAIL_PARSE_PATTERN": "kp0040.php?__kgrnr={committee_id:d}",
"ORGANIZATION_DETAIL_PARSE_PATTERN_FULL": "kp0040.php?__cwp=1&__kgrnr={committee_id:d}",
"ORGANIZATION_DETAIL_PRINT_PATTERN": "%skp0040.php?__kgrnr=%d",
"ORGANIZATION_DETAIL_PRINT_PATTERN_FULL": "%skp0040.php?__cwpall=1&__kgrnr=%d",
"PAPER_DETAIL_PARSE_PATTERN": "vo0050.php?__kvonr={paper_id:d}",
"PAPER_DETAIL_PRINT_PATTERN": "%svo0050.php?__kvonr=%d",
"PERSON_DETAIL_PARSE_PATTERN": "kp0051.php?__kpenr={person_id:d}",
"PERSON_DETAIL_PARSE_PATTERN_ALT": "kp0050.php?__kpenr={person_id:d}&grnr=0",
"PERSON_DETAIL_PRINT_PATTERN": "%skp0051.php?__kpenr=%d",
"PERSON_DETAIL_PRINT_PATTERN_ALT": "%skp0050.php?__kpenr=%d&grnr=0",
"PERSON_ORGANIZATION_PARSE_PATTERN": "kp0050.php?__cwpall=1&__kpenr={person_id:d}",
"PERSON_ORGANIZATION_PRINT_PATTERN": "%skp0050.php?__cwpall=1&__kpenr=%d",
"PERSON_OVERVIEW_PARSE_PATTERN": "kp0041.php?__cwpall=1",
"PERSON_OVERVIEW_PRINT_PATTERN": "%skp0041.php?__cwpall=1"
},
"xpath": {
"MEETING_DETAIL_AGENDAITEM_ROWS": "//*[@class='smccontenttable smc_page_to0040_contenttable']/tbody/tr",
"MEETING_DETAIL_AGENDAITEM_ROWS_PAPER_LINK": ".//a",
"MEETING_DETAIL_FILES": "//*[@id='smccontent']//table",
"MEETING_DETAIL_FILES_CONTAINER_CLASSNAME": "smcdocbox",
"MEETING_DETAIL_IDENTIFIER_TD": "//*[@id=\"smctablevorgang\"]/tbody//td",
"MEETING_DETAIL_ORGANIZATION_LINK": "//li[@class='smcmenucontext_fct_gremium']/a",
"MEETING_DETAIL_TITLE": "//h1",
"PAPER_DETAIL_AGENDA_ROWS": "//*[@class='smccontenttable smc_page_vo0050_contenttable']/tbody/tr",
"PAPER_DETAIL_FILES": "//*[@id='smccontent']//table",
"PAPER_DETAIL_FILES_CONTAINER_CLASSNAME": "smcdocbox",
"PAPER_DETAIL_IDENTIFIER_TD": "//*[@id='smctablevorgang']/tbody//td",
"PAPER_DETAIL_TITLE": "//h1",
"PERSONLIST_LINES": "//table[@id='smc_page_kp0041_contenttable1']//tr",
"PERSON_ORGANIZATION_LINES": "//table[@id='smc_page_kp0050_contenttable1']//tr"
}
},
"user_agent_name": "OParl Scraper 0,1",
"wait_time": 0.2
},
"sitemap_folder": "/srv/www/ris-web/webapp/static/sitemap",
"stopwords_path": "/srv/www/ris-web/config/synonyms_global.txt",
"synonyms_path": "/srv/www/ris-web/config/stopwords_de_global.txt",
"thumbs_path": "/srv/www/ris-web/webapp/static/thumbs/",
"thumbs_sizes": [
800,
300,
150
],
"thumbs_suffix": "jpg",
"thumbs_url": "/static/thumbs",
"thumbs_valid_types": [
"jpg",
"pdf",
"tif",
"bmp",
"png",
"gif"
]
}
]