Skip to content

Commit 869c44c

Browse files
committed
Better vente de regexes - see HEA-833
1 parent e2e7f5b commit 869c44c

2 files changed

Lines changed: 83 additions & 44 deletions

File tree

pipelines/assets/livelihood_activity_regexes.json

Lines changed: 35 additions & 41 deletions
Original file line numberDiff line numberDiff line change
@@ -55,18 +55,6 @@
5555
true,
5656
null
5757
],
58-
[
59-
"vente (?:de |d')?{product_pattern}",
60-
null,
61-
true,
62-
null
63-
],
64-
[
65-
"{product_pattern} sales",
66-
null,
67-
true,
68-
null
69-
],
7058
[
7159
"(?:autre auto-emploi|other self-employment){separator_pattern} {age_gender_pattern}{separator_pattern} {product_pattern}",
7260
"OtherCashIncome",
@@ -314,16 +302,10 @@
314302
"price"
315303
],
316304
[
317-
"prix d'achat \\(espèces\\)",
318-
null,
319-
false,
320-
"purchase_price"
321-
],
322-
[
323-
"prix de vente \\(espèces\\)",
305+
"prix (?:d'|du|de) ?(?:vente|achat) ?\\(?(?:cash|espèces|cash prix d'achat)\\)?",
324306
null,
325307
false,
326-
"sales_price"
308+
"price"
327309
],
328310
[
329311
"paiement en {unit_of_measure_pattern} par fois \\({product_pattern}\\)",
@@ -500,7 +482,25 @@
500482
"percentage_sold_slaughtered"
501483
],
502484
[
503-
"{product_pattern} sales: no. sold",
485+
"{product_pattern} sales{separator_pattern} {nbr_pattern} sold",
486+
"LivestockSale",
487+
true,
488+
"quantity_sold"
489+
],
490+
[
491+
"vente (?:des? |d')?{product_pattern}{separator_pattern} {nbr_pattern} {vendu_pattern}",
492+
"LivestockSale",
493+
true,
494+
"quantity_sold"
495+
],
496+
[
497+
"{product_pattern} vente: {nbr_pattern} {vendu_pattern}",
498+
"LivestockSale",
499+
true,
500+
"quantity_sold"
501+
],
502+
[
503+
"{product_pattern}: {nbr_pattern} {vendu_pattern}",
504504
"LivestockSale",
505505
true,
506506
"quantity_sold"
@@ -745,24 +745,6 @@
745745
false,
746746
"carcass_weight"
747747
],
748-
[
749-
"vente (?:des? |d')?{product_pattern}: {nbr_pattern} {vendu_pattern}",
750-
"LivestockSale",
751-
true,
752-
"quantity_sold"
753-
],
754-
[
755-
"{product_pattern} vente: {nbr_pattern} {vendu_pattern}",
756-
"LivestockSale",
757-
true,
758-
"quantity_sold"
759-
],
760-
[
761-
"{product_pattern}: {nbr_pattern} {vendu_pattern}",
762-
"LivestockSale",
763-
true,
764-
"quantity_sold"
765-
],
766748
[
767749
"autre (?:culture de rente|culture|cereale|céréale/légumineuse|céréale? principal|légumineuse)?{separator_pattern} \\(?{product_pattern}\\)?{separator_pattern} {unit_of_measure_pattern} produits?",
768750
null,
@@ -902,7 +884,7 @@
902884
"quantity_sold"
903885
],
904886
[
905-
"{product_pattern}: (?P<unit_of_measure_id>kg|l)? {vendu_pattern}",
887+
"(?:vente de |vente )?{product_pattern}{separator_pattern} (?P<unit_of_measure_id>kg|l)? {vendu_pattern}",
906888
null,
907889
true,
908890
"quantity_sold"
@@ -919,14 +901,26 @@
919901
true,
920902
"quantity_sold"
921903
],
904+
[
905+
"(?:autre)?{separator_pattern}vente (?:de |d')?{product_pattern}",
906+
null,
907+
true,
908+
null
909+
],
910+
[
911+
"{product_pattern} sales",
912+
null,
913+
true,
914+
null
915+
],
922916
[
923917
"autre achat{separator_pattern} ?{product_pattern}",
924918
null,
925919
true,
926920
null
927921
],
928922
[
929-
"autre auto-emploi{separator_pattern} ?{product_pattern}",
923+
"autre auto-emploi{separator_pattern} (?:vente de |vente )?{product_pattern}",
930924
null,
931925
true,
932926
null

pipelines_tests/test_assets/test_livelihood_activity_regexes.json

Lines changed: 48 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -165,6 +165,18 @@
165165
"product_id": "eggs",
166166
"attribute": "quantity_sold"
167167
},
168+
"vente de légume: kg vendus": {
169+
"is_start": true,
170+
"product_id": "légume",
171+
"unit_of_measure_id": "kg",
172+
"attribute": "quantity_sold"
173+
},
174+
"vente cultures maraichères: kg vendus": {
175+
"is_start": true,
176+
"product_id": "cultures maraichères",
177+
"unit_of_measure_id": "kg",
178+
"attribute": "quantity_sold"
179+
},
168180
"Vente de bois de chauffage": {
169181
"is_start": true,
170182
"product_id": "bois de chauffage"
@@ -185,6 +197,10 @@
185197
"is_start": true,
186198
"product_id": "orange"
187199
},
200+
"autre: vente néré": {
201+
"is_start": true,
202+
"product_id": "néré"
203+
},
188204
"sand sales": {
189205
"is_start": true,
190206
"product_id": "sand"
@@ -530,6 +546,18 @@
530546
"product_id": "volaille",
531547
"attribute": "quantity_sold"
532548
},
549+
"vente de porcs - nb vendus": {
550+
"strategy_type": "LivestockSale",
551+
"is_start": true,
552+
"product_id": "porcs",
553+
"attribute": "quantity_sold"
554+
},
555+
"pig sales: no sold": {
556+
"strategy_type": "LivestockSale",
557+
"is_start": true,
558+
"product_id": "pig",
559+
"attribute": "quantity_sold"
560+
},
533561
"Peaux: nb vendues": {
534562
"strategy_type": "LivestockSale",
535563
"is_start": true,
@@ -874,10 +902,19 @@
874902
"attribute": "price"
875903
},
876904
"prix d'achat (espèces)": {
877-
"attribute": "purchase_price"
905+
"attribute": "price"
878906
},
879907
"prix de vente (espèces)": {
880-
"attribute": "sales_price"
908+
"attribute": "price"
909+
},
910+
"prix du vente (cash prix d'achat)": {
911+
"attribute": "price"
912+
},
913+
"prix du vente (cash)": {
914+
"attribute": "price"
915+
},
916+
"prix de vente (cash)": {
917+
"attribute": "price"
881918
},
882919
"dépense": {
883920
"attribute": "expenditure"
@@ -1088,7 +1125,15 @@
10881125
"is_start": true,
10891126
"product_id": "petty trade",
10901127
"attribute": "people_per_household"
1091-
},
1128+
},
1129+
"autre auto-emploi vente de foin": {
1130+
"is_start": true,
1131+
"product_id": "foin"
1132+
},
1133+
"autre auto-emploi: vente prods de cueillette": {
1134+
"is_start": true,
1135+
"product_id": "prods de cueillette"
1136+
},
10921137
"hommes - preparation terre/labour": {
10931138
"is_start": true,
10941139
"product_id": "preparation terre/labour",

0 commit comments

Comments
 (0)