From bb0cc64638f4ed4858b7f42d4cd91733e8f6203a Mon Sep 17 00:00:00 2001
From: Mali Akmanalp <mali@akmanalp.com>
Date: Mon, 6 Mar 2017 17:33:16 -0500
Subject: [PATCH] Add livestock classification for colombia rural.

---
 product/Datlas/Rural/README.md                |   3 ++
 product/Datlas/Rural/Tupfile                  |   3 ++
 product/Datlas/Rural/clean_livestock.py       |  34 ++++++++++++++++++
 .../Datlas/Rural/in/Livestock - Hierarchy.tsv |   9 +++++
 product/Datlas/Rural/in/Livestock - Names.tsv |  10 ++++++
 product/Datlas/Rural/out/livestock.csv        |  10 ++++++
 product/Datlas/Rural/out/livestock.dta        | Bin 0 -> 2156 bytes
 product/Datlas/Rural/raw/livestock_dept.dta   | Bin 0 -> 11796 bytes
 setup.py                                      |   3 +-
 9 files changed, 71 insertions(+), 1 deletion(-)
 create mode 100644 product/Datlas/Rural/README.md
 create mode 100644 product/Datlas/Rural/Tupfile
 create mode 100644 product/Datlas/Rural/clean_livestock.py
 create mode 100644 product/Datlas/Rural/in/Livestock - Hierarchy.tsv
 create mode 100644 product/Datlas/Rural/in/Livestock - Names.tsv
 create mode 100644 product/Datlas/Rural/out/livestock.csv
 create mode 100644 product/Datlas/Rural/out/livestock.dta
 create mode 100644 product/Datlas/Rural/raw/livestock_dept.dta

diff --git a/product/Datlas/Rural/README.md b/product/Datlas/Rural/README.md
new file mode 100644
index 0000000..96ff5ee
--- /dev/null
+++ b/product/Datlas/Rural/README.md
@@ -0,0 +1,3 @@
+This is basically HS92, with names generously translated by Ryan.
+
+[here](https://docs.google.com/spreadsheets/d/1yLhTZ6cV6aWLJkJmAQGRIEGKFSEeF_dwMtCA3YagOzI/edit#gid=0)
diff --git a/product/Datlas/Rural/Tupfile b/product/Datlas/Rural/Tupfile
new file mode 100644
index 0000000..706d2e7
--- /dev/null
+++ b/product/Datlas/Rural/Tupfile
@@ -0,0 +1,3 @@
+CLEAN = PYTHONPATH=../../../ python2.7 -B clean_livestock.py
+
+: in/* |> $(CLEAN) |> out/livestock.csv out/livestock.dta
diff --git a/product/Datlas/Rural/clean_livestock.py b/product/Datlas/Rural/clean_livestock.py
new file mode 100644
index 0000000..0597e93
--- /dev/null
+++ b/product/Datlas/Rural/clean_livestock.py
@@ -0,0 +1,34 @@
+import pandas as pd
+
+from classification import (Hierarchy, repeated_table_to_parent_id_table,
+                            parent_code_table_to_parent_id_table,
+                            Classification)
+
+if __name__ == "__main__":
+    names = pd.read_table("./in/Livestock - Names.tsv", encoding="utf-8",
+                          dtype={"code": str})
+
+    hierarchy = pd.read_table("./in/Livestock - Hierarchy.tsv", encoding="utf-8")
+    hierarchy.columns = ["level1_code", "level0_code"]
+
+    fields = {
+        "level0": [],
+        "level1": []
+    }
+
+    h = Hierarchy(["level0", "level1"])
+    parent_code_table = repeated_table_to_parent_id_table(hierarchy, h, fields)
+    parent_code_table.code = parent_code_table.code.astype(str)
+
+    parent_code_table = parent_code_table.merge(names, on=["code", "level"])
+
+    parent_id_table = parent_code_table_to_parent_id_table(parent_code_table, h)
+    parent_id_table["name"] = parent_id_table.name_en
+
+    parent_id_table = parent_id_table[["code", "name", "level", "name_en",
+                                       "name_es", "parent_id"]]
+
+    c = Classification(parent_id_table, h)
+
+    c.to_csv("out/livestock.csv")
+    c.to_stata("out/livestock.dta")
diff --git a/product/Datlas/Rural/in/Livestock - Hierarchy.tsv b/product/Datlas/Rural/in/Livestock - Hierarchy.tsv
new file mode 100644
index 0000000..cd3de97
--- /dev/null
+++ b/product/Datlas/Rural/in/Livestock - Hierarchy.tsv	
@@ -0,0 +1,9 @@
+level1	level0
+buffalos	total
+equines	total
+fisheries	total
+fishing	total
+goats	total
+pigs	total
+poultry	total
+sheep	total
\ No newline at end of file
diff --git a/product/Datlas/Rural/in/Livestock - Names.tsv b/product/Datlas/Rural/in/Livestock - Names.tsv
new file mode 100644
index 0000000..a10c47d
--- /dev/null
+++ b/product/Datlas/Rural/in/Livestock - Names.tsv	
@@ -0,0 +1,10 @@
+code	level	name_en	name_es
+total	level0	Total	Total
+buffalos	level1	Buffaloes	Búfalos
+equines	level1	Equines	Equinos
+fisheries	level1	Fisheries	Fisheries
+fishing	level1	Fishing	Fishing
+goats	level1	Goats	Goats
+pigs	level1	Pigs	Pigs
+poultry	level1	Poultry	Poultry
+sheep	level1	Sheep	Sheep
\ No newline at end of file
diff --git a/product/Datlas/Rural/out/livestock.csv b/product/Datlas/Rural/out/livestock.csv
new file mode 100644
index 0000000..0797f5d
--- /dev/null
+++ b/product/Datlas/Rural/out/livestock.csv
@@ -0,0 +1,10 @@
+"","code","name","level","name_en","name_es","parent_id"
+0,"total","Total","level0","Total","Total",""
+1,"buffalos","Buffaloes","level1","Buffaloes","Búfalos",0.0
+2,"equines","Equines","level1","Equines","Equinos",0.0
+3,"fisheries","Fisheries","level1","Fisheries","Fisheries",0.0
+4,"fishing","Fishing","level1","Fishing","Fishing",0.0
+5,"goats","Goats","level1","Goats","Goats",0.0
+6,"pigs","Pigs","level1","Pigs","Pigs",0.0
+7,"poultry","Poultry","level1","Poultry","Poultry",0.0
+8,"sheep","Sheep","level1","Sheep","Sheep",0.0
diff --git a/product/Datlas/Rural/out/livestock.dta b/product/Datlas/Rural/out/livestock.dta
new file mode 100644
index 0000000000000000000000000000000000000000..dfb1c88004e658ed8ba558c10a53247328e337bd
GIT binary patch
literal 2156
zcmeH{!3u&v5QeuPsJHOaU5%1D2GuDL>K<E-wy^D*t3hwysziyiR)@%gBgoJ0|ILge
zOx*2|zJ7iX_QYD#*m7;hvm9?Wa>+nPPsc$XVaMSL+`;P#dj7xw!ki?}0LpB5_F%v3
zq0aE=0rQl4zzo&G!>Ots)}1sF7=QZ>%>(L2-rltnXeaQ83GCA#pj;|4F9fJpk>8L+
zCX{c>FSw*kfL=-qT<DA_7cg~#n1NcF1y(QPm<S2UID^2rP|_q{oh!4zxQK<t#`Vnt
Y<z&i}_(q6t78s|e0#Ww=H)erxcSSSE(EtDd

literal 0
HcmV?d00001

diff --git a/product/Datlas/Rural/raw/livestock_dept.dta b/product/Datlas/Rural/raw/livestock_dept.dta
new file mode 100644
index 0000000000000000000000000000000000000000..69b6c18e3a7817a24c7f378eb43b643253cf553e
GIT binary patch
literal 11796
zcmd6td303u700hp6g@6caj8nr6suybBoni8ro$6~Dp490E4XDQ$pi)_3o`=(YW=w&
z)QUz?5wtoiA}DICdO$_bGos)gkRx(%85UP6Mn$A<>G$`3H_4NC=N)QK&*|?E9y9sO
zUB37JmiJ~tsdO|Qt*%Q)!=bUUXk9EB4kcrWSTq$22ZLpyg02^#nkng6Q&IzGpFb>A
z(DO1>8SX`)g354cMEK;r2>%@s4ke;Bu|ydEH~$Hx;|(#mYG@808Yl`3i`4{*3WH^V
zVCleMd8nZ0uc3mjC>o;8Vd61|uYF)%k3oF>7hjLbGw=7?+g!Kvbx|+#`iA|?_1#m=
zbsJxM4=}I$@HJGRy@n=5lksRxB37NA(i}^L^JstCmv(mU><ks`_9u-tMv+=L(Nr5v
z$D10f<8@RQYmO$<4Y9^FHKn5onnEd}M0^qyXsR7giP)rAg79AL`0Dz2YHYl53@uo&
zpc|-WfN}zj6B}w`Nz&G;$23LLDY$@z`e?EtRn32d3&))saN;-dhEh$*bRwQghxZ~@
zi*y;g-wxF`B^zKmd_qNv#u%OU>|gVNTADjYk531k=8U%nnNN2NGVdQ`nd_H>=ofro
zK%ow4Pd&^%q2fRMK&YTAeUvycF;?yLakbHIcNtO#<i(C6Qm4@+r*lUVWxwBKIiOl~
z2GMWMpzQV`q%IlK{o(&HpliK*;60t~r1P_`t#)-SQRIugbtauWKHhj%o_^hjqu=Y3
z^2<7}COU}%&Y-u8Hm$*Hn|h|f|EWPUJpxVjf$j-Mn5e%z<sO61{DT7H|2u=ORffwH
zsvR3mMr+ftBqnUUHWkJj;n9d$FWgjmZ{HvHBh_58b)&99mal<Qygn~>esx?IEpqG!
z`j43vrf^Lso#7NT<XX4f1q&6xgZ46PPrXdSf|S$Cqya@(TiE2;duno=Z{XeQX}4iL
zxAsVpRN5!jU{9py?5V*~am1Qa;;u-m*)oLKt(<k(gu{g;X=l4AJKby%h49`!X~$nr
z*D_Qcxzj2vX>OXBNGGRYH_>}Y5bvM)RzCv-+BX4+O_&&OjOp#B3!uoyotCy>hu;HW
zwu_$uYMku=wD>jvsj;zGGxigg1^VzQ1JFbrLwpQA{xb{HlZY8Zz4tnRu073dfwE{m
zjJihw)J&|ek0zQ@F2Fj9SIfxNYyrI<OZ8-dzBU$)zTYY=H$r`584Fq{8rpV+69eTv
z13=J0o!e$<klvjSz^kOSWHDQ~UAxd(<cc911t1AEHDg@RdD?|nsE(%*vmWYJ075OO
zb+G|5wH$*!9k_(*)*%uHASg*AXBi7>s1*zvp%9g$q!vp3cBW;7O0)nFgrf3kY@q>w
zpAJ+AKuD7Ycj<5u07<C)7Isk&K+-`~8i2|<hFm((U;vU1l+P9jfVbQ<ZyM`Rb)^+N
z%?S0SODMbe$5D<@_9<Hd`03Dm1ppxn^~OpQL*yE^kSmL?<^k{qY4<<d#6TUd01%4R
z9RA9<IQ(@0f)1)?wE@&b075O)jy412D;<NkX|tV>MUPu;E68iqPcEU_`_7-}2$j9#
z8vud@`&;lBD)Kx4KP3xq1|XyvICrySf#U4|ge)RMcQ}CR1)-!BhFazr3slQ7<f_H(
zm>1GIX>I(_Sa@O?0B^CZx{FwcBhfHrrla<DE#~TkA4WPtS>LY%;HQI{2|!S?eGLri
zaqxox1fl5E4JOE&p3)XP`-Oe|X(k4`vJHTsq&k2bCNd6y4l<WetFHqfv>XjzYGMdP
z00{Gf5_cFNldyukm~V4+6dn9!KQr|@-v!{!g0icB?Evc1l>np|7CmHKbff@CF`UUu
z5RC^QD4AXVkz+wk&@TLRIPN|GLa`!!k(3_W{sBOcnj&qM{yXKP?R|Hg$W>Ob!gSPk
zE}@#24xHfV(2?8;KuYtS+l&h<20%#Dn!nWnYzsYE;n$O^?*t&EX<ven^qBD^03l8F
z<MRec%mpA7Ysjqz2sCRKo~CUcaJWg6PWcRgPc76E+bSwFLjBVv6y3AzY)7c9odw`^
zL7n?onvPnw8i16=h|f%r>3abPN@l*>?f|OSivWCrq)1A~;9Y40d5Bq$QBSjlTqlq~
z41l1ddX1&dCIATikX_6Ntutp^MZD|vcCS&-?o;FF(A>HMfDq)-CxORMz1IT}T-e)r
z32MDrTkurVR)2^(76Qea0SH-GhitL53+lKMfS;0U=QEH?s395<iq)~=N8@5zI{=~F
z_Mo>7Kr^f&-t~@k33XWez6+cz=+dtM2rj75?3*gK27sXC6PsaBkL(Hnerdil3xFWC
z>hq!ra^Zd2g-7b&UGU9($Ay}{1%Mzm4dVl6APqp$;r;iGi?YW7=pb`ReKb0XcRg&B
z-3Lw@+jqDVL*(vH07ybz^o(&)JP&}7CgH||oKXqDo2DAR>@1_?=C=U&1=)&fihEXe
z47)g8yYSA<Ogr!X)0YDfgd*#Fw$Ptr$hB_{SZ5XU$^2`#P@i8>?POs;{T={Ghsk`d
z&Nc$@>R|0U)L2lf+5t%AZn=qrtOnrMa*-D8LeOC}V%Fo`2LK4wv|oP90R5k37r7c{
z@=O4dl0%jm3+<O%#l_x}IjyW7><HDe;adQLg+6PJFiJkK27sVs)yF*bK7Sj4pAMM}
z06_<PD$nlsU($eQ>d`?ZrA`d0YCQl!sQkPRCr#@~9i(5=s*3?g)f~y2yQ;T}dFr{3
z2kL{V=CBv}Vlx0iDC<D8`_sDF0K8eKEq^}WWbrl9l;&n;Fz;C{_W}^oY^~>psd*WI
zUz*L67|1no>{$RL9cJ?~NSGJW<fZ5Mh_cJJO89v2E7xLfM}8XPq!}sN1i-H!7A*uI
z#js+zY1)<~0HNjRuz3auBmfAtQ0rbZz^EGm2rajA8e-OC<x&lJTF(Cd0Fwn3v;**y
zy6QRrLa_oH%#l?ES6U@S-dmXLL7j0YhL+?H00axPWSx_RI(!uXNvLAJ5ew7+kc29{
z)3HE-aRB_X=yN*&K}kw3H5O<(TgcH-R>>Lwg4DadG51+&>DwCcNNwFW&*-4`-ENie
z8s*;glzrKyj!@R;8vzJH+4fo!<iZsIye{lR4>1<#m{$NuLRIp<8MqdJU(=3)4!$c)
zs{$Y>sSa#2K`x&OKoBbX9B+$59|s_nJI&=*wN@z~IC1pqJ|1Lmy!A@c)%QFKz%L8B
z41nZf#&ulo;Q$1w>G&rNpl$}>mD<`mfn(U`ZVh;My{hN<xY#irfK-dfE$m`606|IW
zKZAp;0^rxxnFg!0)J#3MP?0akHyWX?egS~wVku7?JUlBcm9og=ESBE|K+2+pgT&*<
z(o(-#ShD~KQmcpBOq$pZjD=jgf6FugQZ@6hVWIvAz*{W(=oHprajR9zPY-swI_i?X
zBNsVBwLkbZ06(ci3mM3<K2R2Eicb$>pPL{re;k02h1%L~fXUYa5NZ*56EW+t;Yk33
zlD0DUoA%N*+Ja{ZD*U+8q#3CA5P;ARfdtRZU5{90yzAj9S@+bV(KClSLRn)s0T6=B
z++?1rQE?jpAxH|EAk|eG5OkpOS4^5J4nSx*_3HBmh)f0G*K$>V1K`zxF0zK0c`@=e
zwvemk@D!y?+<Di&U}C76Y?bjE<vvFhcJxU*F4UAy00>#od5eq%`U?PWkgfB#r5p?P
z#`PNTt}u}(A5k_u1;9^-cjhpVE6uvO0Hj(({%(TIHUJPBh7RGT?SBgZq1+j@#j&6w
zuUciiMg?5wC@L9tmy;$vjE<6)px1_j*W)jb0g#kDj+Jar0g$p7dzTXf(WL-{AniN&
z`NMU~bqt<~6Buy32~ur%6M*0Pfag+W(n_o1Z6=1m6>K5b0zxaSa=zF5)IIes%RAB0
zLDjYc@CHc_e%8+jb^kK}{Q9AAnzrEWsLW6Bs7GKi07=Oi_ZomM1R(Xp=*Jjn;23hX
z98CZq)FM*4#aIYF0YEBu<9q|yF{_*(v;0h`#$qEB4R{}bl*O5k7z<f#L2}W-TQiLU
z;HM<r3_z&GOMOi()WDb7LM|Qb82}_D4}Qr6S$K_h;Vo9Al8bfnG5~^-t;d?XMRh})
zRnAin2gshqy!^hP(=b-$dH_Ne8-8y9wRV*TyfIj3@QJi{27shwd(vb<=Kv6bq*I<V
zz>p;X{G?v;I0LzqTz(4xsoZL+aiPuvAW6NDAChFoS>=2@P;Q~r7jLqKMgW2i)B;`)
zHADlR1%&9~Hw<7`0q|Qus5by9NE&D?5V3_^3y6I&07+_f4qF%iKq?l68R!E*Qj&T`
z1+P(9T6;Qb*O7&0395a{st^|t%qU}{9{@q93lCun^$!4$vY^3i;Y0v_N~#M0NNK7r
z3;8+*Z&%Y~-Y;;jt`Ju_YdptL&N1Yw#ni=C1>fsk>lP~ips$RDjgJBFYx^Aw0m!xB
z7&-=3NH=0c{ak>QX69N0s1Y1Pu9-!UTDl>n?phQ9kfg?!@#Ku&0EBW!wwPyP)cYQ*
zf){h#dAT}j-(F5fsUO+^Nb_PWZ%LVY0Dk?DX=Wf-vGQjFkfgRJa27)W2w8lxEzh+5
zz()ZHO{<RJBkT1Gv<0E%?&SN;y4wIqO5%pxD2cA-2X_g#Q0fVOh`J6PB|W$!Ww8YU
zAY@@Tc3HR^fD{AucUkb0n*Ia}eqDVJx>|bl*VD9^o8oghJPUJlu+IY^70dan`KN-v
z0A1x4D!bPFx=9Ux!K(26`V!q;A$+$;E%kixp1PDR<m#v}`a$}BlR|vojqRX9`i_;x
z@V2OK1R&J{cSoEg07-}J4Ge?;2rZ{(K^8rVXIY^F=l5hG{LC$Z-wGD+)%AN&{qAlr
IdwvN1Z;{4l9smFU

literal 0
HcmV?d00001

diff --git a/setup.py b/setup.py
index 3ff9765..9f0eef3 100644
--- a/setup.py
+++ b/setup.py
@@ -2,7 +2,7 @@
 
 setup(
     name="linnaeus",
-    version="v0.0.68",
+    version="v0.0.69",
     author="Mali Akmanalp <Harvard CID>",
     description=("Harvard CID's classification tools."),
     url="http://github.com/cid-harvard/classifications/",
@@ -20,6 +20,7 @@
             'industry/ISIC/Colombia_Prosperia/out/industries_colombia_isic_prosperia.csv',
             'industry/NAICS/Mexico/out/industries_mexico_scian_2007.csv',
             'industry/NAICS/Mexico_datlas/out/industries_mexico_scian_2007_datlas.csv',
+            'product/Datlas/Rural/out/livestock.csv',
             'product/HS/Atlas/out/hs92_atlas.csv',
             'product/HS/Mexico_Prospedia/out/products_mexico_prospedia.csv',
             'product/HS/Colombia_Prospedia/out/products_colombia_prospedia.csv',