From d37cd49f7d3b765d4c74cd182a9ed91118f07f69 Mon Sep 17 00:00:00 2001
From: virio-andreyana <arsyan.virio@gmail.com>
Date: Wed, 14 Dec 2022 12:19:44 +0100
Subject: [PATCH] add script for retrieving lng terminals

---
 Snakefile                               |  5 ++-
 scripts/build_gas_input_locations.py    | 26 +------------
 scripts/retrieve_gas_input_locations.py | 50 +++++++++++++++++++++++++
 3 files changed, 55 insertions(+), 26 deletions(-)
 create mode 100644 scripts/retrieve_gas_input_locations.py

diff --git a/Snakefile b/Snakefile
index 468783e1..4011dee0 100644
--- a/Snakefile
+++ b/Snakefile
@@ -128,10 +128,13 @@ if config["sector"]["gas_network"] or config["sector"]["H2_retrofit"]:
         resources: mem_mb=4000
         script: "scripts/build_gas_network.py"
 
+    rule retrieve_gas_input_locations:
+        output: "data/gas_network/Europe-Gas-Tracker-August-2022.geojson"
+        script: 'scripts/retrieve_gas_input_locations.py'
 
     rule build_gas_input_locations:
         input:
-            lng="data/gas_network/Europe-Gas-Tracker-August-2022.xlsx",
+            lng="data/gas_network/Europe-Gas-Tracker-August-2022.geojson",
             entry="data/gas_network/scigrid-gas/data/IGGIELGN_BorderPoints.geojson",
             production="data/gas_network/scigrid-gas/data/IGGIELGN_Productions.geojson",
             regions_onshore=pypsaeur("resources/regions_onshore_elec_s{simpl}_{clusters}.geojson"),
diff --git a/scripts/build_gas_input_locations.py b/scripts/build_gas_input_locations.py
index 22529193..c6e3d634 100644
--- a/scripts/build_gas_input_locations.py
+++ b/scripts/build_gas_input_locations.py
@@ -18,34 +18,10 @@ def read_scigrid_gas(fn):
     df.drop(["param", "uncertainty", "method"], axis=1, inplace=True)
     return df
 
-def read_lng_gem(fn):
-    df = pd.read_excel(fn, sheet_name = 'LNG terminals - data')
-    df = df.set_index("ComboID")
-
-    remove_status = ['Cancelled']
-    remove_country = ['Cyprus','Turkey']
-    remove_terminal = ['Puerto de la Luz LNG Terminal','Gran Canaria LNG Terminal']
-
-    for index in df.index:
-        if df.Status[index] in remove_status:
-            df = df.drop([index])
-        elif df.Country[index] in remove_country:
-            df = df.drop([index])
-        elif df.TerminalName[index] in remove_terminal:
-            df = df.drop([index])
-        elif df.CapacityInMtpa[index] == "--":
-            df = df.drop([index])
-        else:
-            continue
-    
-    geometry = gpd.points_from_xy(df['Longitude'], df['Latitude'])
-    gdf = gpd.GeoDataFrame(df, geometry=geometry, crs="EPSG:4326")
-    return gdf
-
 def build_gas_input_locations(lng_fn, entry_fn, prod_fn, countries):
     
     # LNG terminals
-    lng = read_lng_gem(lng_fn)
+    lng = gpd.read_file(lng_fn)
 
     # Entry points from outside the model scope
     entry = read_scigrid_gas(entry_fn)
diff --git a/scripts/retrieve_gas_input_locations.py b/scripts/retrieve_gas_input_locations.py
new file mode 100644
index 00000000..82840196
--- /dev/null
+++ b/scripts/retrieve_gas_input_locations.py
@@ -0,0 +1,50 @@
+"""
+Retrieve gas infrastructure data fro
+"""
+
+import logging
+from helper import progress_retrieve
+
+import zipfile
+from pathlib import Path
+
+logger = logging.getLogger(__name__)
+
+
+if __name__ == "__main__":
+    if 'snakemake' not in globals():
+        from helper import mock_snakemake
+        snakemake = mock_snakemake('retrieve_gas_network_data')
+        rootpath = '..'
+    else:
+        rootpath = '.'
+
+# LNG terminals
+
+    lng_url="https://globalenergymonitor.org/wp-content/uploads/2022/09/Europe-Gas-Tracker-August-2022.xlsx",
+    
+    storage_options = {'User-Agent': 'Mozilla/5.0'}
+    df = pd.read_excel(lng_url, storage_options=storage_options, sheet_name = 'LNG terminals - data')
+    #df = pd.read_excel(fn, sheet_name = 'LNG terminals - data')
+    df = df.set_index("ComboID")
+
+    remove_status = ['Cancelled']
+    remove_country = ['Cyprus','Turkey']
+    remove_terminal = ['Puerto de la Luz LNG Terminal','Gran Canaria LNG Terminal']
+
+    for index in df.index:
+        if df.Status[index] in remove_status:
+            df = df.drop([index])
+        elif df.Country[index] in remove_country:
+            df = df.drop([index])
+        elif df.TerminalName[index] in remove_terminal:
+            df = df.drop([index])
+        elif df.CapacityInMtpa[index] == "--":
+            df = df.drop([index])
+        else:
+            continue
+    
+    geometry = gpd.points_from_xy(df['Longitude'], df['Latitude'])
+    lng = gpd.GeoDataFrame(df, geometry=geometry, crs="EPSG:4326")
+
+    lng.to_file(snakemake.output[0])
\ No newline at end of file