Update peru location classification to include XX99 locations and new

callao PERU-122

location/Peru/datlas/clean.py

@@ -1,8 +1,6 @@
 import pandas as pd
 
-from classification import (Hierarchy, parent_code_table_to_parent_id_table,
-                            Classification)
-
+from classification import (Hierarchy, Classification)
 
 if __name__ == "__main__":
 
@@ -19,10 +17,57 @@
     # Drop the "district" / level since we don't use it
     df = df[df.level != "district"]
 
+    def get_id_by_code(df, code):
+        rows = df[df.code == code].index.tolist()
+        assert len(rows) == 1
+        return float(rows[0])
+
+    # Create the "Other" provinces that are XX9900 for unknown province data
+    # points
+    def create_others(row):
+        row.parent_id = get_id_by_code(df, row.code)
+        row.code = row.code[:2] + "9900"
+        row.level = "msa"
+        row.name_en = row.name_en + " (unknown province)"
+        row.name_short_en = row.name_short_en + " (unknown province)"
+        row.name_es = row.name_es + " (provincia desconocida)"
+        row.name_short_es = row.name_short_es + " (provincia desconocida)"
+        row.name = row.name_en
+        return row
+
+    others = df[df.level == "department"].apply(create_others, axis=1)
+
+    # Add Callao province that moved from Callo dept that no longer exists
+    callao = pd.Series({
+        "code": "159800",
+        "level": "msa",
+        "name_es": u"Callao",
+        "name_short_es": u"Callao",
+        "name_en": u"Callao",
+        "name_short_en": u"Callao",
+        "name": u"Callao",
+        "parent_id": get_id_by_code(df, "150000")  # Callao's parent is Lima
+    })
+
+    df = pd.concat([df,
+                    others,
+                    pd.DataFrame(callao).T,
+                    ]).reset_index(drop=True)
+
+    df.parent_id = df.parent_id.astype(float)
+
     h = Hierarchy(["country", "department", "msa", "municipality"])
     df.level = df.level.astype("category", categories=h, ordered=True)
     df.level = df.level.astype(str)
 
+    # Drop old Callao department and province
+    # Do this after reset_index to not mess up the id order
+    df = df[df.code != "070000"]
+    df = df[df.code != "070100"]
+
+    # Order the columns
+    df = df[["code","name","level","name_es","name_en","name_short_es","name_short_en","parent_id"]]
+
     c = Classification(df, h)
 
     c.to_csv("out/locations_peru_datlas.csv")

location/Peru/datlas/out/locations_peru_datlas.csv

@@ -6,7 +6,6 @@
 4,"040000","Arequipa","department","Arequipa","Arequipa","Arequipa","Arequipa",0.0
 5,"050000","Ayacucho","department","Ayacucho","Ayacucho","Ayacucho","Ayacucho",0.0
 6,"060000","Cajamarca","department","Cajamarca","Cajamarca","Cajamarca","Cajamarca",0.0
-7,"070000","Callao","department","Callao","Callao","Callao","Callao",0.0
 8,"080000","Cusco","department","Cusco","Cusco","Cusco","Cusco",0.0
 9,"090000","Huancavelica","department","Huancavelica","Huancavelica","Huancavelica","Huancavelica",0.0
 10,"100000","Huanuco","department","Huanuco","Huanuco","Huanuco","Huanuco",0.0
@@ -91,7 +90,6 @@
 89,"061100","San Miguel","msa","San Miguel","San Miguel","San Miguel","San Miguel",6.0
 90,"061200","San Pablo","msa","San Pablo","San Pablo","San Pablo","San Pablo",6.0
 91,"061300","Santa Cruz","msa","Santa Cruz","Santa Cruz","Santa Cruz","Santa Cruz",6.0
-92,"070100","Callao","msa","Callao","Callao","Callao","Callao",7.0
 93,"080100","Cusco","msa","Cusco","Cusco","Cusco","Cusco",8.0
 94,"080200","Acomayo","msa","Acomayo","Acomayo","Acomayo","Acomayo",8.0
 95,"080300","Anta","msa","Anta","Anta","Anta","Anta",8.0
@@ -220,3 +218,29 @@
 218,"250200","Atalaya","msa","Atalaya","Atalaya","Atalaya","Atalaya",25.0
 219,"250300","Padre Abad","msa","Padre Abad","Padre Abad","Padre Abad","Padre Abad",25.0
 220,"250400","Purus","msa","Purus","Purus","Purus","Purus",25.0
+221,"019900","Amazonas","msa","Amazonas (provincia desconocida)","Amazonas (unknown province)","Amazonas (provincia desconocida)","Amazonas (unknown province)",1.0
+222,"029900","Ancash","msa","Ancash (provincia desconocida)","Ancash (unknown province)","Ancash (provincia desconocida)","Ancash (unknown province)",2.0
+223,"039900","Apurimac","msa","Apurimac (provincia desconocida)","Apurimac (unknown province)","Apurimac (provincia desconocida)","Apurimac (unknown province)",3.0
+224,"049900","Arequipa","msa","Arequipa (provincia desconocida)","Arequipa (unknown province)","Arequipa (provincia desconocida)","Arequipa (unknown province)",4.0
+225,"059900","Ayacucho","msa","Ayacucho (provincia desconocida)","Ayacucho (unknown province)","Ayacucho (provincia desconocida)","Ayacucho (unknown province)",5.0
+226,"069900","Cajamarca","msa","Cajamarca (provincia desconocida)","Cajamarca (unknown province)","Cajamarca (provincia desconocida)","Cajamarca (unknown province)",6.0
+227,"079900","Callao","msa","Callao (provincia desconocida)","Callao (unknown province)","Callao (provincia desconocida)","Callao (unknown province)",7.0
+228,"089900","Cusco","msa","Cusco (provincia desconocida)","Cusco (unknown province)","Cusco (provincia desconocida)","Cusco (unknown province)",8.0
+229,"099900","Huancavelica","msa","Huancavelica (provincia desconocida)","Huancavelica (unknown province)","Huancavelica (provincia desconocida)","Huancavelica (unknown province)",9.0
+230,"109900","Huanuco","msa","Huanuco (provincia desconocida)","Huanuco (unknown province)","Huanuco (provincia desconocida)","Huanuco (unknown province)",10.0
+231,"119900","Ica","msa","Ica (provincia desconocida)","Ica (unknown province)","Ica (provincia desconocida)","Ica (unknown province)",11.0
+232,"129900","Junin","msa","Junin (provincia desconocida)","Junin (unknown province)","Junin (provincia desconocida)","Junin (unknown province)",12.0
+233,"139900","La Libertad","msa","La Libertad (provincia desconocida)","La Libertad (unknown province)","La Libertad (provincia desconocida)","La Libertad (unknown province)",13.0
+234,"149900","Lambayeque","msa","Lambayeque (provincia desconocida)","Lambayeque (unknown province)","Lambayeque (provincia desconocida)","Lambayeque (unknown province)",14.0
+235,"159900","Lima","msa","Lima (provincia desconocida)","Lima (unknown province)","Lima (provincia desconocida)","Lima (unknown province)",15.0
+236,"169900","Loreto","msa","Loreto (provincia desconocida)","Loreto (unknown province)","Loreto (provincia desconocida)","Loreto (unknown province)",16.0
+237,"179900","Madre de Dios","msa","Madre de Dios (provincia desconocida)","Madre de Dios (unknown province)","Madre de Dios (provincia desconocida)","Madre de Dios (unknown province)",17.0
+238,"189900","Moquegua","msa","Moquegua (provincia desconocida)","Moquegua (unknown province)","Moquegua (provincia desconocida)","Moquegua (unknown province)",18.0
+239,"199900","Pasco","msa","Pasco (provincia desconocida)","Pasco (unknown province)","Pasco (provincia desconocida)","Pasco (unknown province)",19.0
+240,"209900","Piura","msa","Piura (provincia desconocida)","Piura (unknown province)","Piura (provincia desconocida)","Piura (unknown province)",20.0
+241,"219900","Puno","msa","Puno (provincia desconocida)","Puno (unknown province)","Puno (provincia desconocida)","Puno (unknown province)",21.0
+242,"229900","San Martin","msa","San Martin (provincia desconocida)","San Martin (unknown province)","San Martin (provincia desconocida)","San Martin (unknown province)",22.0
+243,"239900","Tacna","msa","Tacna (provincia desconocida)","Tacna (unknown province)","Tacna (provincia desconocida)","Tacna (unknown province)",23.0
+244,"249900","Tumbes","msa","Tumbes (provincia desconocida)","Tumbes (unknown province)","Tumbes (provincia desconocida)","Tumbes (unknown province)",24.0
+245,"259900","Ucayali","msa","Ucayali (provincia desconocida)","Ucayali (unknown province)","Ucayali (provincia desconocida)","Ucayali (unknown province)",25.0
+246,"159800","Callao","msa","Callao","Callao","Callao","Callao",15.0

setup.py

@@ -2,7 +2,7 @@
 
 setup(
     name="linnaeus",
-    version="v0.0.57",
+    version="v0.0.58",
     author="Mali Akmanalp <Harvard CID>",
     description=("Harvard CID's classification tools."),
     url="http://github.com/cid-harvard/classifications/",