arabesque import tests

2 files changed, 10 insertions, 0 deletions

diff --git a/python/tests/files/arabesque_example.json b/python/tests/files/arabesque_example.json
new file mode 100644
index 00000000..50d96b15
--- /dev/null
+++ b/python/tests/files/arabesque_example.json
@@ -0,0 +1,10 @@
+{"postproc_status": "200", "final_sha1": "XXLYXZKYAC5RZGS6I4AFXLC6IESHSPD3", "final_status_code": "200", "final_domain": "files.eccomasproceedia.org", "hit": 1, "breadcrumbs": "-", "identifier": "10.123/abc", "final_timestamp": "201810300511", "final_was_dedupe": 0, "initial_domain": "files.eccomasproceedia.org", "final_url": "https://files.eccomasproceedia.org/papers/compdyn-2015/1114.pdf?mtime=20170329172028", "initial_url": "https://files.eccomasproceedia.org/papers/compdyn-2015/1114.pdf?mtime=20170329172028", "final_mimetype": "application/pdf"}
+{"postproc_status": "500", "final_sha1": "X7WDYGM7CTCNP232F5TWK2A62RQDFQX3", "final_status_code": "200", "final_domain": "oru.diva-portal.org", "hit": 1, "breadcrumbs": "-", "identifier": "10.123/abc", "final_timestamp": "201810300511", "final_was_dedupe": 0, "initial_domain": "oru.diva-portal.org", "final_url": "http://oru.diva-portal.org/smash/get/diva2:138559/FULLTEXT01", "initial_url": "http://oru.diva-portal.org/smash/get/diva2:138559/FULLTEXT01", "final_mimetype": "application/pdf"}
+{"postproc_status": null, "final_sha1": "L763R3EZCHPH7XJLNRHBYXL5Q2LAEJC5", "final_status_code": "200", "final_domain": "files.eccomasproceedia.org", "hit": 1, "breadcrumbs": "-", "identifier": "10.7712/120113.4766.c1181", "final_timestamp": "201810300511", "final_was_dedupe": 0, "initial_domain": "files.eccomasproceedia.org", "final_url": "https://files.eccomasproceedia.org/papers/compdyn-2013/C1181.pdf?mtime=20170330153901", "initial_url": "https://files.eccomasproceedia.org/papers/compdyn-2013/C1181.pdf?mtime=20170330153901", "final_mimetype": "application/pdf"}
+{"postproc_status": null, "final_sha1": "KZR7WSB2G4IMKE2Z6ON5SN6E3AI7LYQZ", "final_status_code": "200", "final_domain": "www.termedia.pl", "hit": 1, "breadcrumbs": "-", "identifier": null, "final_timestamp": "201810300511", "final_was_dedupe": 1, "initial_domain": "www.termedia.pl", "final_url": "https://www.termedia.pl/Journal/-41/pdf-15163-7?filename=Trudnosci", "initial_url": "https://www.termedia.pl/Journal/-41/pdf-15163-7?filename=Trudnosci", "final_mimetype": "application/pdf"}
+{"postproc_status": null, "final_sha1": "YAKODW2FUCFLAHJIVIQ55ZJTLWRI5IDY", "final_status_code": "200", "final_domain": "www.scielo.br", "hit": 1, "breadcrumbs": "-", "identifier": "10.1590/s0104-11691993000200003", "final_timestamp": "201810300511", "final_was_dedupe": 1, "initial_domain": "www.scielo.br", "final_url": "http://www.scielo.br/pdf/rlae/v1n2/v1n2a03.pdf", "initial_url": "http://www.scielo.br/pdf/rlae/v1n2/v1n2a03.pdf", "final_mimetype": "application/pdf"}
+{"postproc_status": null, "final_sha1": "W6A5C3W45ZPSKUKINNEOKHF7VIGLNT7V", "final_status_code": "200", "final_domain": "files.eccomasproceedia.org", "hit": 1, "breadcrumbs": "-", "identifier": "10.7712/100016.2436.6566", "final_timestamp": "201810300511", "final_was_dedupe": 0, "initial_domain": "files.eccomasproceedia.org", "final_url": "https://files.eccomasproceedia.org/papers/eccomas-congress-2016/6566.pdf?mtime=20170308164900", "initial_url": "https://files.eccomasproceedia.org/papers/eccomas-congress-2016/6566.pdf?mtime=20170308164900", "final_mimetype": "application/pdf"}
+{"postproc_status": null, "final_sha1": "L5ZJCCH5NIV2EDOTQ54M5VIAYSVQWSNN", "final_status_code": "200", "final_domain": "files.eccomasproceedia.org", "hit": 1, "breadcrumbs": "-", "identifier": "10.7712/120117.5608.18010", "final_timestamp": "201810300511", "final_was_dedupe": 0, "initial_domain": "files.eccomasproceedia.org", "final_url": "https://files.eccomasproceedia.org/papers/compdyn-2017/18010.pdf?mtime=20171002111627", "initial_url": "https://files.eccomasproceedia.org/papers/compdyn-2017/18010.pdf?mtime=20171002111627", "final_mimetype": "application/pdf"}
+{"postproc_status": null, "final_sha1": "DHPTE53TRRLZSAP4SO3A3BL5AAVWIT26", "final_status_code": "200", "final_domain": "files.eccomasproceedia.org", "hit": 1, "breadcrumbs": "-", "identifier": "10.7712/100016.2360.4678", "final_timestamp": "201810300511", "final_was_dedupe": 0, "initial_domain": "files.eccomasproceedia.org", "final_url": "https://files.eccomasproceedia.org/papers/eccomas-congress-2016/4678.pdf?mtime=20170308164728", "initial_url": "https://files.eccomasproceedia.org/papers/eccomas-congress-2016/4678.pdf?mtime=20170308164728", "final_mimetype": "application/pdf"}
+{"postproc_status": null, "final_sha1": "FM5ZQWTUQ2N7T7SXFNLCVA6N5RWQRTI6", "final_status_code": "200", "final_domain": "files.eccomasproceedia.org", "hit": 1, "breadcrumbs": "-", "identifier": "10.7712/100016.2380.8613", "final_timestamp": "201810300511", "final_was_dedupe": 0, "initial_domain": "files.eccomasproceedia.org", "final_url": "https://files.eccomasproceedia.org/papers/eccomas-congress-2016/8613.pdf?mtime=20170308165111", "initial_url": "https://files.eccomasproceedia.org/papers/eccomas-congress-2016/8613.pdf?mtime=20170308165111", "final_mimetype": "application/pdf"}
+{"postproc_status": null, "final_sha1": "Z3KITYKYIOYXXXFRUMVXVU6WS4VPWE3D", "final_status_code": "200", "final_domain": "www.scielo.br", "hit": 1, "breadcrumbs": "-", "identifier": "10.1590/s1413-99362009000200012", "final_timestamp": "201810300511", "final_was_dedupe": 1, "initial_domain": "www.scielo.br", "final_url": "http://www.scielo.br/pdf/pci/v14n2/v14n2a12.pdf", "initial_url": "http://www.scielo.br/pdf/pci/v14n2/v14n2a12.pdf", "final_mimetype": "application/pdf"}
diff --git a/python/tests/files/arabesque_example.sqlite3 b/python/tests/files/arabesque_example.sqlite3
new file mode 100644
index 00000000..7c89bff0
--- /dev/null
+++ b/python/tests/files/arabesque_example.sqlite3