Merge pull request #3 from VishwamAI/update-ci-workflow

Update ci workflow

.github/workflows/ci.yml

@@ -37,17 +37,18 @@ jobs:
     - name: Run linting
       run: flake8 .
 
-    - name: Run tests
-      run: |
-        pip install pytest pytest-cov
-        pytest --cov=src
+    # Tests have been removed as per user request
+    # - name: Run tests
+    #   run: |
+    #     pip install pytest pytest-cov
+    #     pytest --cov=src
 
-    - name: Upload coverage results
-      if: success()
-      uses: actions/upload-artifact@v3
-      with:
-        name: coverage-report
-        path: coverage.xml
+    # - name: Upload coverage results
+    #   if: success()
+    #   uses: actions/upload-artifact@v3
+    #   with:
+    #     name: coverage-report
+    #     path: coverage.xml
 
     - name: Debug information
       run: |

src/agent_management.py

@@ -7,7 +7,7 @@
 import jax.numpy as jnp
 import tensorflow as tf
 from src.trading_environment import TradingEnvironment
-from moving_average_crossover_strategy import crossover_strategy_jax, crossover_strategy_tf
+from moving_average_crossover_strategy import moving_average
 
 class TradingAgent:
     def __init__(self, entity: Entity, ledger_api: LedgerApi, contract: Contract):
@@ -22,16 +22,16 @@ def make_decision(self, observation):
         # Extract relevant information from the observation
         balance, shares_held, short_ma, long_ma = observation
 
-        # Use both JAX and TensorFlow crossover strategies to generate signals
+        # Use JAX-based moving average crossover strategy to generate signal
         price = self.environment.prices[self.environment.current_step]
         prices_jax = jnp.array([price])
-        prices_tf = tf.constant([price])
 
-        signal_jax = crossover_strategy_jax(prices_jax, self.short_window, self.long_window)[-1]
-        signal_tf = crossover_strategy_tf(prices_tf, self.short_window, self.long_window)[-1].numpy()
+        short_ma = moving_average(prices_jax, self.short_window)[-1]
+        long_ma = moving_average(prices_jax, self.long_window)[-1]
+        signal = jnp.where(short_ma > long_ma, 1, -1)
 
-        # Combine signals (e.g., take the average)
-        combined_signal = (signal_jax + signal_tf) / 2
+        # Use the signal directly for decision making
+        combined_signal = signal
 
         # Convert combined signal to action
         if combined_signal > 0.5 and balance > 0:

src/trading_environment.py

@@ -2,7 +2,7 @@
 import numpy as np
 import pandas as pd
 from gym import spaces
-from moving_average_crossover_strategy import moving_average_jax, load_data
+from moving_average_crossover_strategy import moving_average as moving_average_jax, load_data
 
 class TradingEnvironment(gym.Env):
     def __init__(self, csv_file_path, initial_balance=10000, transaction_fee=0.001):