fix

biomap/checkpoint/model/model.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
 oid sha256:106fe1ea7f4f0819e360823374bce7840a1a150b39a2e45090612c159a25dfca
-size 95521785
+size 95521785

biomap/helper.py

@@ -15,7 +15,7 @@ import streamlit as st
 import cv2
 
 @st.cache_data(hash_funcs={LitUnsupervisedSegmenter: lambda dt: dt.name})
-def inference_on_location(model, longitude=2.98, latitude=48.81, start_date=2020, end_date=2022, how="year"):
+def inference_on_location(model, latitude=48.81, longitude=2.98, start_date=2020, end_date=2022, how="year"):
     """Performe an inference on the latitude and longitude between the start date and the end date
 
     Args:
@@ -65,13 +65,13 @@ def inference_on_location(model, longitude=2.98, latitude=48.81, start_date=2020
     
     images = [np.asarray(img) for img in imgs]
     labeled_imgs = [np.asarray(img) for img in labeled_imgs]
-    title=f"TimeLapse at location {tuple(location)} between {start_date} and {end_date}"
+    title=f"TimeLapse at location ({location[0]:.2f},{location[1]:.2f}) between {start_date} and {end_date}"
     fig = plot_imgs_labels(dates, images, labeled_imgs, scores_details, scores, title=title)
     # fig.save("test.png")
     return fig
 
 @st.cache_data(hash_funcs={LitUnsupervisedSegmenter: lambda dt: dt.name})
-def inference_on_location_and_month(model, longitude = 2.98, latitude = 48.81, start_date = '2020-03-20'):
+def inference_on_location_and_month(model, latitude = 48.81, longitude = 2.98, start_date = '2020-03-20'):
     """Performe an inference on the latitude and longitude between the start date and the end date
 
     Args:
@@ -100,7 +100,7 @@ def inference_on_location_and_month(model, longitude = 2.98, latitude = 48.81, s
     logging.info(f"Calculated Biodiversity Score : {score}")
     img, label, labeled_img = transform_to_pil(outputs[0])
     
-    title=f"Prediction at location {tuple(location)} at {start_date}"
+    title=f"Prediction at location ({location[0]:.2f},{location[1]:.2f})  at {start_date}"
     fig = plot_image([start_date], [np.asarray(img)], [np.asarray(labeled_img)], [score_details], [score],title=title)
     return fig
 

biomap/streamlit_app.py

@@ -54,72 +54,103 @@ def init_app(cfg_name) -> LitUnsupervisedSegmenter:
 def app(model):
     if "infered" not in st.session_state:
         st.session_state["infered"] = False
+    if "submit" not in st.session_state:
+        st.session_state["submit"] = False
+    if "submit2" not in st.session_state:
+        st.session_state["submit2"] = False
 
     st.markdown("<h1 style='text-align: center;'>🐢 Biomap by Ekimetrics 🐢</h1>", unsafe_allow_html=True)
     st.markdown("<h2 style='text-align: center;'>Estimate Biodiversity in the world with the help of land cover.</h2>", unsafe_allow_html=True)
     st.markdown("<p style='text-align: center;'>The segmentation model is an association of UNet and DinoV1 trained on the dataset CORINE. Land use is divided into 6 differents classes : Each class is assigned a GBS score from 0 to 1</p>", unsafe_allow_html=True)
     st.markdown("<p style='text-align: center;'>Buildings : 0.1 | Infrastructure : 0.1 | Cultivation : 0.4 | Wetland : 0.9 | Water : 0.9 | Natural green : 1 </p>", unsafe_allow_html=True)
     st.markdown("<p style='text-align: center;'>The score is then averaged on the full image.</p>", unsafe_allow_html=True)
+    if st.session_state["submit"]:
+            fig = inference_on_location(model, st.session_state["lat"], st.session_state["long"], st.session_state["start_date"], st.session_state["end_date"], st.session_state["segment_interval"])
+            st.session_state["infered"] = True
+            st.session_state["previous_fig"] = fig
     
-    col_1, col_2 = st.columns([0.5,0.5])
-    with col_1:
-        m = folium.Map(location=[DEFAULT_LATITUDE, DEFAULT_LONGITUDE], zoom_start=DEFAULT_ZOOM)
-        m.add_child(folium.LatLngPopup())
-        f_map = st_folium(m, width=FOLIUM_WIDTH, height=FOLIUM_HEIGHT)
-    
-    selected_latitude = DEFAULT_LATITUDE
-    selected_longitude = DEFAULT_LONGITUDE
-
-    if f_map.get("last_clicked"):
-        selected_latitude = f_map["last_clicked"]["lat"]
-        selected_longitude = f_map["last_clicked"]["lng"]
-
-    with col_2:
-        tabs1, tabs2 = st.tabs(["TimeLapse", "Single Image"])
-        with tabs1:
-            col_tab1_1, col_tab1_2 = st.columns(2)
-            with col_tab1_1:
-                lat = st.text_input("latitude", value=selected_latitude)
-            with col_tab1_2:
-                long = st.text_input("longitude", value=selected_longitude)
-
-            col_tab1_11, col_tab1_22 = st.columns(2)
-            years = list(range(MIN_YEAR, MAX_YEAR, 1))
-            with col_tab1_11:
-                start_date = st.selectbox("Start date", years)
-
-            end_years = [year for year in years if year > start_date]
-            with col_tab1_22:
-                end_date = st.selectbox("End date", end_years) 
+    if st.session_state["submit2"]:
+            fig = inference_on_location_and_month(model, st.session_state["lat_2"], st.session_state["long_2"], st.session_state["date_2"])
+            st.session_state["infered"] = True
+            st.session_state["previous_fig"] = fig
+
+    if st.session_state["infered"]:
+            st.plotly_chart(st.session_state["previous_fig"], use_container_width=True)
+
+    m = folium.Map(location=[DEFAULT_LATITUDE, DEFAULT_LONGITUDE], zoom_start=DEFAULT_ZOOM)
+    m.add_child(folium.LatLngPopup())
+    tabs1, tabs2 = st.tabs(["TimeLapse", "Single Image"])
+    with tabs1:
+        
+        
+        submit = st.button("Predict TimeLapse", use_container_width=True, type="primary")
+        st.session_state["submit"] = submit
+        
+        col_1, col_2 = st.columns([0.5,0.5])
+        with col_1:
+            
+            f_map = st_folium(m, key="tab1", width=FOLIUM_WIDTH, height=FOLIUM_HEIGHT)
+        
+        selected_latitude = DEFAULT_LATITUDE
+        selected_longitude = DEFAULT_LONGITUDE
+
+        if f_map.get("last_clicked"):
+            selected_latitude = f_map["last_clicked"]["lat"]
+            selected_longitude = f_map["last_clicked"]["lng"]
+
+        with col_2:
+                col_tab1_1, col_tab1_2 = st.columns(2)
+                with col_tab1_1:
+                    lat = st.text_input("latitude", value=selected_latitude)
+                    st.session_state["lat"] = lat
+                with col_tab1_2:
+                    long = st.text_input("longitude", value=selected_longitude)
+                    st.session_state["long"] = long
+
+                col_tab1_11, col_tab1_22 = st.columns(2)
+                years = list(range(MIN_YEAR, MAX_YEAR, 1))
+                with col_tab1_11:
+                    start_date = st.selectbox("Start date", years)
+                    st.session_state["start_date"] = start_date
+
+                end_years = [year for year in years if year > start_date]
+                with col_tab1_22:
+                    end_date = st.selectbox("End date", end_years) 
+                    st.session_state["end_date"] = end_date
+            
+                segment_interval = st.radio("Interval of time between two segmentation", options=['month','2months', 'year'],horizontal=True)
+                st.session_state["segment_interval"] = segment_interval
+
+    with tabs2:
+        submit2 = st.button("Predict Single Image", use_container_width=True, type="primary")
+        st.session_state["submit2"] = submit2
+
+        col_1_tab_2, col_2_tab_2 = st.columns([0.5,0.5])
+        with col_1_tab_2:
+            m_tab_2 = folium.Map(location=[DEFAULT_LATITUDE, DEFAULT_LONGITUDE], zoom_start=DEFAULT_ZOOM)
+            m_tab_2.add_child(folium.LatLngPopup())
+            f_map_tab_2 = st_folium(m, key="tab2", width=FOLIUM_WIDTH, height=FOLIUM_HEIGHT)
         
-            segment_interval = st.radio("Interval of time between two segmentation", options=['month','2months', 'year'],horizontal=True)
-            submit = st.button("Predict TimeLapse", use_container_width=True)
-        with tabs2:
+        selected_latitude_2 = DEFAULT_LATITUDE
+        selected_longitude_2 = DEFAULT_LONGITUDE
+
+        if f_map_tab_2.get("last_clicked"):
+            selected_latitude_2 = f_map_tab_2["last_clicked"]["lat"]
+            selected_longitude_2 = f_map_tab_2["last_clicked"]["lng"]
+
+        with col_2_tab_2:
             col_tab2_1, col_tab2_2 = st.columns(2)
             with col_tab2_1:
-                lat = st.text_input("lat.", value=selected_latitude)
+                lat_2 = st.text_input("lat.", value=selected_latitude_2)
+                st.session_state["lat_2"] = lat_2
             with col_tab2_2:
-                long = st.text_input("long.", value=selected_longitude)
-
-            date = st.text_input("date", "2021-01-01", placeholder="2021-01-01")
-        
-            submit2 = st.button("Predict Single Image", use_container_width=True)
+                long_2 = st.text_input("long.", value=selected_longitude_2)
+                st.session_state["long_2"] = long_2
 
+            date_2 = st.text_input("date", "2021-01-01", placeholder="2021-01-01")
+            st.session_state["date_2"] = date_2
 
-    if submit:
-        fig = inference_on_location(model, lat, long, start_date, end_date, segment_interval)
-        st.session_state["infered"] = True
-        st.session_state["previous_fig"] = fig
 
-    if submit2:
-        fig = inference_on_location_and_month(model, lat, long, date)
-        st.session_state["infered"] = True
-        st.session_state["previous_fig"] = fig
-
-    if st.session_state["infered"]:
-        st.plotly_chart(st.session_state["previous_fig"], use_container_width=True)
-        
-    
 if __name__ == "__main__":
     model = init_app("my_train_config.yml")
     app(model)

biomap/utils_gee.py

@@ -6,10 +6,16 @@ import matplotlib.pyplot as plt
 import os
 from pathlib import Path
 import logging
+import json
 
 #Initialize
-service_account = 'cvimg-355@cvimg-377115.iam.gserviceaccount.com'
-credentials = ee.ServiceAccountCredentials(service_account, os.path.join(os.path.dirname(__file__), '.private-key.json'))
+service_account = os.environ["SERVICE_ACCOUNT_EE"]
+private_key = json.loads(os.environ["PRIVATE_KEY_EE"])
+
+with open(os.path.join(os.path.dirname(__file__), '.private-key-2.json'), "w") as ipt:
+    json.dump(private_key, ipt)
+
+credentials = ee.ServiceAccountCredentials(service_account, os.path.join(os.path.dirname(__file__), '.private-key-2.json'))
 ee.Initialize(credentials)
 
 def get_image(location, d1, d2):
@@ -143,6 +149,8 @@ def extract_img(location,start_date,end_date, width = 0.01 , len = 0.01,scale=5)
     Returns:
         img: image as numpy array
     """
+    # reversed longitude latitude
+    location = (location[1], location[0])
     ee_img, geometry = extract_ee_img(location, width,start_date,end_date , len)
     url = get_url(ee_img, geometry, scale)
     img = extract_np_from_url(url)