Merge pull request #8 from danyoungday/ci

Fixing Github actions to perform CI

.flake8

@@ -1,2 +1,3 @@
 [flake8]
-max-line-length = 120
+max-line-length = 120
+exclude = inputSpecs.py

.github/workflows/enroads.yml

@@ -24,12 +24,12 @@ jobs:
         if [ -f requirements.txt ]; then pip install -r requirements.txt; fi
     - name: Download En-ROADS sdk
       env:
+        ENROADS_URL: ${{ secrets.ENROADS_URL }}
         ENROADS_ID: ${{ secrets.ENROADS_ID }}
         ENROADS_PASSWORD: ${{ secrets.ENROADS_PASSWORD }}
-        ENROADS_URL: ${{ secrets.ENROADS_URL }}
       run: python -m enroadspy.download_sdk
     - name: Lint with PyLint
-      run: pylint .
+      run: pylint ./*
     - name: Lint with Flake8
       run: flake8
     - name: Run unit tests

.pylintrc

@@ -10,4 +10,4 @@ suggestion-mode=yes
 
 good-names=X,F,X0
 
-fail-under=9.8
+fail-under=9.6

app/components/outcome.py

@@ -1,8 +1,6 @@
 """
 OutcomeComponent class for the outcome section of the app.
 """
-import sys
-
 from dash import Input, Output, html, dcc
 import dash_bootstrap_components as dbc
 import pandas as pd
@@ -26,7 +24,7 @@ def __init__(self, evolution_handler: EvolutionHandler, all_cand_idxs: list[str]
                              "_global_population_in_2100"]
         self.plot_outcomes = ["Temperature change from 1850",
                               "Adjusted cost of energy per GJ",
-                              "Government net revenue from adjustments", 
+                              "Government net revenue from adjustments",
                               "Total Primary Energy Demand"]
 
     def plot_outcome_over_time(self, outcome: str, outcomes_jsonl: list[list[dict[str, float]]], cand_idxs: list[int]):
@@ -36,7 +34,7 @@ def plot_outcome_over_time(self, outcome: str, outcomes_jsonl: list[list[dict[st
         TODO: Fix colors to match parcoords
         """
         outcomes_dfs = [pd.DataFrame(outcomes_json) for outcomes_json in outcomes_jsonl]
-        color_map = [c for c in px.colors.qualitative.Plotly]
+        color_map = px.colors.qualitative.Plotly
 
         fig = go.Figure()
         showlegend = True
@@ -87,7 +85,7 @@ def plot_outcome_over_time(self, outcome: str, outcomes_jsonl: list[list[dict[st
                 "xanchor": "center"},
         )
         return fig
-    
+
     def create_outcomes_div(self):
         """
         Note: We have nested loads here. The outer load is for both graphs and triggers when the outcomes store
@@ -173,7 +171,7 @@ def update_outcomes_store(*context_values):
             outcomes_jsonl = [outcomes_df[self.plot_outcomes].to_dict("records") for outcomes_df in outcomes_dfs]
 
             # Parse energy demand policy
-            colors = ["brown", "red", "blue", "green", "pink", "lightblue", "orange"]
+            # colors = ["brown", "red", "blue", "green", "pink", "lightblue", "orange"]
             energies = ["coal", "oil", "gas", "renew and hydro", "bio", "nuclear", "new tech"]
             demands = [f"Primary energy demand of {energy}" for energy in energies]
             selected_dfs = [outcomes_dfs[i] for i in self.all_cand_idxs[:-2]]

enroadspy/download_sdk.py

@@ -24,10 +24,15 @@ def main():
         url = os.getenv("ENROADS_URL")
         username = os.getenv("ENROADS_ID")
         password = os.getenv("ENROADS_PASSWORD")
+        assert url is not None, \
+            "Please set the ENROADS_URL environment variable."
         assert username is not None and password is not None, \
             "Please set the ENROADS_ID and ENROADS_PASSWORD environment variables. \
             To get access to them go to https://en-roads.climateinteractive.org/ and sign up."
 
+        assert len(url) > 0 and len(username) > 0 and len(password) > 0, \
+            "There was an issue with the environment variables, one or more are empty."
+
         r = requests.get(url, auth=(username, password), timeout=60)
 
         if r.status_code == 200:

enroadspy/enroads_runner.py

@@ -90,7 +90,7 @@ def run_enroads(self, input_str=None):
         En-ROADS main function requires a file as input.
         Possible variable values are stored in inputSpecs.jsonl.
         From the documentation:
-            The input string is a space-delimited list of index-value pairs, where a colon separates the 
+            The input string is a space-delimited list of index-value pairs, where a colon separates the
             index number from the value number with no spaces. Index numbers are zero-based.
         NOTE: The indices are the line numbers in inputSpecs.jsonl starting from 0, NOT the id column.
         """

evolution/candidate.py

@@ -5,7 +5,6 @@
 from pathlib import Path
 
 import numpy as np
-import pandas as pd
 import torch
 
 from enroadspy import load_input_specs

evolution/crossover/crossover.py

@@ -1,18 +1,22 @@
+"""
+Abstract crossover class to be inherited.
+"""
 from abc import ABC, abstractmethod
 
 from evolution.candidate import Candidate
 from evolution.mutation.mutation import Mutation
 
+
 class Crossover(ABC):
     """
     Abstract class for crossover operations.
     """
-    def __init__(self, mutator: Mutation=None):
+    def __init__(self, mutator: Mutation = None):
         self.mutator = mutator
-    
+
     @abstractmethod
     def crossover(self, cand_id: str, parent1: Candidate, parent2: Candidate) -> list[Candidate]:
         """
         Crosses over 2 parents to create offspring. Returns a list so we can return multiple if needed.
         """
-        raise NotImplementedError
+        raise NotImplementedError

evolution/crossover/uniform_crossover.py

@@ -1,20 +1,28 @@
+"""
+Standard uniform crossover
+"""
 import torch
 
 from evolution.candidate import Candidate
 from evolution.crossover.crossover import Crossover
 from evolution.mutation.mutation import Mutation
 
+
 class UniformCrossover(Crossover):
     """
     Crosses over 2 parents.
     We do not keep track of what's in the models and assume they are loaded correct with the parents.
     """
-    def __init__(self, mutator: Mutation=None):
+    def __init__(self, mutator: Mutation = None):
         super().__init__(mutator)
         self.type = "uniform"
 
     def crossover(self, cand_id: str, parent1: Candidate, parent2: Candidate) -> list[Candidate]:
-        child = Candidate(cand_id, [parent1.cand_id, parent2.cand_id], parent1.model_params, parent1.actions, parent1.outcomes)
+        child = Candidate(cand_id,
+                          [parent1.cand_id, parent2.cand_id],
+                          parent1.model_params,
+                          parent1.actions,
+                          parent1.outcomes)
         with torch.no_grad():
             child.model.load_state_dict(parent1.model.state_dict())
             for param, param2 in zip(child.model.parameters(), parent2.model.parameters()):

evolution/evaluation/data.py

@@ -34,7 +34,7 @@ def generate_new_zero_carbon_df(self) -> pd.DataFrame:
             "_new_tech_breakthrough_setting": [0, 1, 2]
         }
         return pd.DataFrame(data)
-    
+
     def generate_renewable_df(self, context: list[str], n=10, seed=42) -> pd.DataFrame:
         """
         Generates our renewables breakthrough context.
@@ -47,7 +47,7 @@ def generate_renewable_df(self, context: list[str], n=10, seed=42) -> pd.DataFra
             min_val = input_specs[input_specs["varId"] == col]["minValue"].iloc[0]
             max_val = input_specs[input_specs["varId"] == col]["maxValue"].iloc[0]
             data[col] = [default_val] + rng.uniform(min_val, max_val, n-1).tolist()
-        
+
         data_df = pd.DataFrame(data)
         return data_df
 
@@ -62,7 +62,7 @@ def __init__(self, context: list[str]):
         if context == ["_new_tech_breakthrough_setting"]:
             df = self.generate_new_zero_carbon_df()
             self.context_vals = df[context].values
-        
+
         elif context == breakthrough_cols:
             df = self.generate_renewable_df(context)
             self.context_vals = df[context].values
@@ -79,7 +79,7 @@ def __init__(self, context: list[str]):
             else:
                 df[col] = scaler.fit_transform(df[col].values.reshape(-1, 1))
         self.tensor_context = torch.tensor(df.values, dtype=torch.float32)
-        
+
     def __getitem__(self, idx):
         return self.tensor_context[idx], self.context_vals[idx]
 

evolution/evaluation/evaluator.py

@@ -19,8 +19,7 @@ class Evaluator:
     Evaluates candidates by generating the actions and running the enroads model on them.
     Generates and stores context data based on config using ContextDataset.
     """
-    def __init__(self, context: list[str], actions: list[str], outcomes: dict[str, bool]):        
-
+    def __init__(self, context: list[str], actions: list[str], outcomes: dict[str, bool]):
         self.actions = actions
         self.outcomes = outcomes
         self.outcome_manager = OutcomeManager(outcomes)

evolution/mutation/mutation.py

@@ -1,7 +1,11 @@
+"""
+Abstract mutation class to implement.
+"""
 from abc import ABC, abstractmethod
 
 import torch
 
+
 class Mutation(ABC):
     """
     Abstract class handling mutation.
@@ -31,5 +35,3 @@ def mutate_(self, candidate):
         Mutate model in-place
         """
         raise NotImplementedError
-
-

evolution/mutation/uniform_mutation.py

@@ -1,9 +1,13 @@
+"""
+Standard uniform mutation implementation.
+"""
 from evolution.candidate import Candidate
 from evolution.mutation.mutation import Mutation
 
+
 class UniformMutation(Mutation):
     """
-    Uniformly mutates 
+    Uniformly mutates
     """
     def __init__(self, mutation_factor, mutation_rate):
         super().__init__(mutation_factor, mutation_rate)
@@ -14,4 +18,4 @@ def mutate_(self, candidate: Candidate):
         Mutate model with gaussian percentage in-place
         """
         for param in candidate.model.parameters():
-            self.gaussian_pct_(param.data)
+            self.gaussian_pct_(param.data)

evolution/outcomes/average_cost.py

@@ -1,13 +1,19 @@
+"""
+Implementation of the AverageCostOutcome class.
+"""
 import pandas as pd
 
 from evolution.outcomes.outcome import Outcome
 
 
 class AverageCostOutcome(Outcome):
+    """
+    Average cost after decision point year.
+    """
     def process_outcomes(self, _, outcomes_df: pd.DataFrame) -> float:
         """
         Returns the average cost from 2025 onwards.
         """
         cost_col = outcomes_df["Adjusted cost of energy per GJ"]
         cost = cost_col.iloc[2025-1990:].mean()
-        return cost
+        return cost

evolution/outcomes/cost_change_year.py

@@ -1,9 +1,15 @@
+"""
+Outcome that checks the first year the cost of energy changes by more than a given threshold.
+"""
 import pandas as pd
 
 from evolution.outcomes.outcome import Outcome
 
 
 class CostChangeYearOutcome(Outcome):
+    """
+    Cost change year outcome implementation
+    """
     def process_outcomes(self, _, outcomes_df: pd.DataFrame) -> float:
         """
         Returns the first year the cost of energy changes by more than a given threshold (2 default)
@@ -18,4 +24,4 @@ def process_outcomes(self, _, outcomes_df: pd.DataFrame) -> float:
             year = 2101
         else:
             year = greater_years.index[0] + 1990
-        return year
+        return year

evolution/outcomes/energy_change.py

@@ -1,9 +1,15 @@
+"""
+Energy change outcome implementation
+"""
 import pandas as pd
 
 from evolution.outcomes.outcome import Outcome
 
 
 class EnergyChangeOutcome(Outcome):
+    """
+    Checks the percent the distribution changed in a given year averaged over the whole outcomes_df.
+    """
     def __init__(self):
         # We don't want fossil fuels because it double counts
         energies = ["bio", "coal", "gas", "oil", "renew and hydro", "new tech", "nuclear"]

evolution/outcomes/enroads.py

@@ -1,13 +1,20 @@
+"""
+Basic enroads outcome
+"""
 import pandas as pd
 
 from evolution.outcomes.outcome import Outcome
 
+
 class EnroadsOutcome(Outcome):
+    """
+    Just checks the final year of an enroads outcome column.
+    """
     def __init__(self, outcome: str):
         self.outcome = outcome
 
     def process_outcomes(self, _, outcomes_df: pd.DataFrame) -> float:
         """
         Simple case where we return the specified outcome in 2100.
         """
-        return outcomes_df[self.outcome].iloc[-1]
+        return outcomes_df[self.outcome].iloc[-1]

evolution/outcomes/gdp_mse.py

@@ -1,7 +1,11 @@
+"""
+This outcome is not used in our evolution but rather is how we get the correct slider values to match the ar6 ssps.
+"""
 import pandas as pd
 
 from evolution.outcomes.outcome import Outcome
 
+
 class GDPOutcome(Outcome):
     """
     Outcome used in finding the context values to match the scenarios in the AR6 database for the 5 scenarios.
@@ -26,4 +30,4 @@ def process_outcomes(self, _, outcomes_df: pd.DataFrame) -> float:
         gdp_col = outcomes_df["Global GDP"]
         sim_gdp = (gdp_col.iloc[self.year_idxs] * 1000).values
         mse = ((sim_gdp - self.label_gdp) ** 2).mean()
-        return mse
+        return mse

evolution/outcomes/max_cost.py

@@ -1,12 +1,19 @@
+"""
+Max cost outcome implementation.
+"""
 import pandas as pd
 
 from evolution.outcomes.outcome import Outcome
 
+
 class MaxCostOutcome(Outcome):
+    """
+    Gets the max cost over the entire range of years.
+    """
     def process_outcomes(self, _, outcomes_df: pd.DataFrame) -> float:
         """
         Returns the max cost of energy.
         """
         cost_col = outcomes_df["Adjusted cost of energy per GJ"]
         cost = cost_col.max()
-        return cost
+        return cost

evolution/outcomes/near_cost.py

@@ -1,12 +1,19 @@
+"""
+Near cost outcome implementation.
+"""
 import pandas as pd
 
 from evolution.outcomes.outcome import Outcome
 
+
 class NearCostOutcome(Outcome):
+    """
+    Gets the average cost over the next 10 years.
+    """
     def process_outcomes(self, _, outcomes_df: pd.DataFrame) -> float:
         """
         Returns the average cost of energy over the next 10 years.
         """
         cost_col = outcomes_df["Adjusted cost of energy per GJ"]
         cost = cost_col.iloc[2025-1990:2035-1990].mean()
-        return cost
+        return cost