added plots for total jobs done and unique proteins folded

api.py

@@ -8,7 +8,7 @@ import utils
 import pandas as pd
 import uvicorn
 
-from classes import Productivity, ProductivityData, Throughput
+from classes import Productivity, ProductivityData, Last24hProductivityData, Throughput
 
 
 # Global variables (saves time on loading data)
@@ -64,9 +64,9 @@ def productivity_metrics():
     # Unpack the metrics using the correct keys
     result = utils.get_productivity(df_all=data_all, df_24h=data_24h)
     all_time = ProductivityData(**result['all_time'])
-    last_24h = ProductivityData(**result['last_24h'])
+    last_24h = Last24hProductivityData(**result['last_24h'])
 
-    return Productivity(all_time=all_time, last_24h=last_24h)
+    return {"all_time": all_time, "last_24h": last_24h}
 
 
 @app.get("/throughput", response_model=Throughput)

app.py

@@ -31,39 +31,56 @@ st.subheader('Productivity overview')
 st.info('Productivity metrics show how many proteins have been folded, which is the primary goal of the subnet. Metrics are estimated using weights and biases data combined with heuristics.')
 
 productivity_all = requests.get(f'{BASE_URL}/productivity').json()
-productivity = productivity_all['all_time']
+completed_jobs = productivity_all['all_time']['total_completed_jobs']
+
 productivity_24h = productivity_all['last_24h']
+completed_jobs = pd.DataFrame(completed_jobs)
+completed_jobs['last_event_at'] = pd.to_datetime(completed_jobs['updated_at'])
 
-# st.write(productivity_all)
-# # st.write(productivity)
-# st.write(productivity_24h)
+unique_folded = completed_jobs.drop_duplicates(subset=['pdb_id'], keep='first')
+unique_folded['last_event_at'] = pd.to_datetime(unique_folded['updated_at'])
 
 m1, m2 = st.columns(2)
+m1.metric('Unique proteins folded', f'{len(unique_folded):,.0f}', delta=f'{productivity_24h["unique_folded"]:,.0f} (24h)')
+m2.metric('Total jobs completed', f'{len(completed_jobs):,.0f}', delta=f'{productivity_24h["total_completed_jobs"]:,.0f} (24h)')
 
-m1.metric('Unique proteins folded', f'{productivity.get("unique_folded", 0):,.0f}', delta=f'{productivity_24h.get("unique_folded", 0):,.0f} (24h)')
-m2.metric('Total jobs completed', f'{productivity.get("total_completed_jobs", 0):,.0f}', delta=f'{productivity_24h.get("total_completed_jobs", 0):,.0f} (24h)')
+st.markdown('<br>', unsafe_allow_html=True)
 
-# m3.metric('Total simulation steps', f'{productivity.get("total_md_steps"):,.0f}', delta=f'{productivity_24h.get("total_md_steps"):,.0f} (24h)')
 
-# st.markdown('<br>', unsafe_allow_html=True)
 
-# time_binned_data = df.set_index('last_event_at').groupby(pd.Grouper(freq='12h'))
+# time_binned_data_complete = completed_jobs.set_index('last_event_at').groupby(pd.Grouper(freq='12h'))
+# time_binned_data_unique = unique_folded.set_index('last_event_at').groupby(pd.Grouper(freq='12h'))
 
-# PROD_CHOICES = {
-#     'Unique proteins folded': 'unique_pdbs',
-#     'Total simulations': 'total_pdbs',
-#     'Total simulation steps': 'total_md_steps',
-# }
-# prod_choice_label = st.radio('Select productivity metric', list(PROD_CHOICES.keys()), index=0, horizontal=True)
-# prod_choice = PROD_CHOICES[prod_choice_label]
-# steps_running_total = time_binned_data[prod_choice].sum().cumsum()
-# st.plotly_chart(
-#     # add fillgradient to make it easier to see the trend
-#     px.area(steps_running_total, y=prod_choice, 
-#             labels={'last_event_at':'', prod_choice: prod_choice_label},
-#     ).update_traces(fill='tozeroy'),
-#     use_container_width=True,
-# )
+PROD_CHOICES = {
+    'Unique proteins folded': 'unique_pdbs',
+    'Total simulations': 'total_pdbs',
+}
+
+
+
+prod_choice_label = st.radio('Select productivity metric', list(PROD_CHOICES.keys()), index=0, horizontal=True)
+prod_choice = PROD_CHOICES[prod_choice_label]
+
+PROD_DATA = {
+    'unique_pdbs': unique_folded,
+    'total_pdbs': completed_jobs,
+}
+df = PROD_DATA[prod_choice]
+
+df = df.sort_values(by='last_event_at').reset_index()
+
+# Create a cumulative count column
+df['cumulative_jobs'] = df.index + 1
+
+# Plot the cumulative jobs over time
+
+st.plotly_chart(
+    # add fillgradient to make it easier to see the trend
+    px.line(df, x='last_event_at', y='cumulative_jobs', 
+              title='Total Jobs Completed Over Time',
+              labels={'last_event_at': 'Time', 'cumulative_jobs': 'Total Jobs Completed'}).update_traces(fill='tozeroy'),
+    use_container_width=True,
+)
 
 st.markdown('<br>', unsafe_allow_html=True)
 

classes.py

@@ -1,12 +1,17 @@
 from pydantic import BaseModel
 
 class ProductivityData(BaseModel):
+    total_completed_jobs: dict[str, dict[int, str]]
+    
+    
+    
+class Last24hProductivityData(BaseModel):
     unique_folded: int
     total_completed_jobs: int
     
 class Productivity(BaseModel):
     all_time: ProductivityData
-    last_24h: ProductivityData
+    last_24h: Last24hProductivityData
     
 class ThroughputData(BaseModel):
     validator_sent: float

utils.py

@@ -164,7 +164,6 @@ def get_data_transferred(df, unit='GB'):
 def get_productivity(df_all, df_24h): 
     result = {
         'all_time': {
-            'unique_folded': 0,
             'total_completed_jobs': 0
         },
         'last_24h': {
@@ -173,19 +172,16 @@ def get_productivity(df_all, df_24h):
         }
     }
     if df_all is not None:
-        unique_folded_all = len(df_all.pdb_id.value_counts())
-        completed_jobs_all = len(df_all[df_all.active == False])
-
-        total_historical_run_updates = df_all.active.isna().sum() 
-        total_historical_completed_jobs = total_historical_run_updates//10 # this is an estimate based on minimum number of updates per pdb    
+  
+        
+        completed_jobs = df_all[df_all['updated_count'] == 10]
         
         result['all_time'].update({
-            'unique_folded': unique_folded_all,
-            'total_completed_jobs': (completed_jobs_all + total_historical_completed_jobs).item(),
+            'total_completed_jobs': completed_jobs[["updated_at", "pdb_id"]].to_dict(),
         })
         
     if df_24h is not None:
-        completed_jobs_24h = df_24h[df_24h['updated_count'] >= 10]
+        completed_jobs_24h = df_24h[df_24h['updated_count'] == 10]
         unique_completed_jobs_24h = completed_jobs_24h.drop_duplicates(subset=['pdb_id'], keep='first')
         result['last_24h'].update({
             'unique_folded': len(unique_completed_jobs_24h),