From 9a9b3d9ce2551536a6bbd10b0bfc0dd3d372eeca Mon Sep 17 00:00:00 2001
From: lizgzil <lizgzil@hotmail.com>
Date: Mon, 23 Dec 2024 16:24:23 +0000
Subject: [PATCH] Add details for new soc coding index to config

---
 nlp_link/soc_mapper/config.yaml | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

diff --git a/nlp_link/soc_mapper/config.yaml b/nlp_link/soc_mapper/config.yaml
index 95e8346..c1b74a6 100644
--- a/nlp_link/soc_mapper/config.yaml
+++ b/nlp_link/soc_mapper/config.yaml
@@ -1,4 +1,4 @@
-soc_data:
+soc_data_feb_24:
   soc_dir: "s3://nesta-open-data/soc_mapper/soc_coding_index_soc2020_volume2_22022024/soc2020volume2thecodingindexexcel22022024.xlsx"
   sheet_name: "SOC2020 coding index"
   soc_2020_ext_col: "SOC 2020 ext"
@@ -9,6 +9,17 @@ soc_data:
   ug_col: "SOC2020 unit group title"
   add_col: "ADD"
   ind_col: "IND"
+soc_data:
+  soc_dir: "s3://nesta-open-data/soc_mapper/soc2020volume2thecodingindexexcel16102024/soc2020volume2thecodingindexexcel16102024.xlsx"
+  sheet_name: "SOC2020 coding index"
+  soc_2020_ext_col: "SOC_2020_ext"
+  soc_2020_col: "SOC_2020"
+  soc_2010_col: "SOC_2010"
+  natural_order_col: "INDEXOCC_-_natural_word_order"
+  sug_col: "SOC2020_ext_SUG_title"
+  ug_col: "SOC2020_unit_group_title"
+  add_col: "ADD"
+  ind_col: "IND"
 soc_mapper:
   reset_embeddings: False
   match_top_n: 10