prepare assume-framework for first release (#177)

This prepares the first release 0.1.0 of assume-framework

.github/workflows/lint-pytest.yml

@@ -37,15 +37,15 @@ jobs:
         # stop the build if there are Python syntax errors or undefined names
         black --check assume examples tests
         isort --check assume examples tests
-    - name: Test with pytest
+    - name: Test with pytest & integration tests
       run: |
-        pytest --cov --cov-report=xml --junitxml="result.xml" -m "not require_learning"
+        pytest -m "not require_learning"
     - name: Install torch dependencies
       run: |
         python -m pip install -e .[learning]
-    - name: Run Learning with assume cli in in-memory db
+    - name: Test pytest with torch & integration tests
       run: |
-        assume -s example_01_rl -db "sqlite:///"
+        pytest --cov --cov-report=xml --junitxml="result.xml"
 
     - name: Upload tests results
       uses: actions/upload-artifact@v3

README.md

@@ -8,7 +8,7 @@ focus on the German market setup. As a newly developed open-source
 model its priority is to ensure usability and customisability
 for a variety of users and use cases in the energy system modelling community.
 
-The unique feature of the ASSUME tool-box is the integration of **deep reinforcement
+The unique feature of the ASSUME toolbox is the integration of **deep reinforcement
 learning** methods into the behavioural strategies of market agents.
 The model comes with different predefined agent representations for the demand and
 generation side that can be used as plug-and-play modules,
@@ -17,7 +17,7 @@ This setup enables a research of new market designs and dynamics in the energy m
 
 The project is developed by [developers](https://assume.readthedocs.io/en/latest/developers.html) from INATECH at University of Freiburg, IISM at KIT, Fraunhofer ISI, FH Aachen.
 The project ASSUME is funded by the Federal Ministry for Economic
-Affairs and Energy (BMWK).
+Affairs and Climate Action (BMWK).
 
 Documentation
 =============
@@ -30,6 +30,23 @@ Documentation
 Installation
 ============
 
+
+Using pip
+---------
+
+The project is available on [PyPi](https://pypi.org/) and can be simply installed:
+
+```
+pip install assume-framework
+```
+
+if you want to install it with testin capabilities run:
+
+```
+pip install assume-framework[test]
+```
+
+
 Using conda
 -----------
 

assume/common/outputs.py

@@ -41,6 +41,7 @@ def __init__(
         db_engine=None,
         export_csv_path: str = "",
         save_frequency_hours: int = 24,
+        learning_mode: bool = False,
     ):
         super().__init__()
 
@@ -55,13 +56,14 @@ def __init__(
             shutil.rmtree(self.p, ignore_errors=True)
             self.p.mkdir(parents=True)
         self.db = db_engine
+        self.learning_mode = learning_mode
 
         # learning
-        episode = self.simulation_id.split("_")[-1]
-        if episode.isdigit():
-            self.episode = int(episode)
-        else:
-            self.episode = None
+        self.episode = None
+        if self.learning_mode:
+            episode = self.simulation_id.split("_")[-1]
+            if episode.isdigit():
+                self.episode = int(episode)
 
         # contruct all timeframe under which hourly values are written to excel and db
         self.start = start
@@ -339,7 +341,7 @@ async def on_stop(self):
             logger.error(f"No scenario run Yet {e}")
 
     def learning_queries(self):
-        if not self.episode:
+        if not self.learning_mode:
             return []
 
         queries = [

assume/common/scenario_loader.py

@@ -262,7 +262,7 @@ async def load_scenario_folder_async(
 
     index = pd.date_range(
         start=start,
-        end=end + pd.Timedelta(hours=24),
+        end=end,
         freq=config["time_step"],
     )
     # get extra parameters for bidding strategies

assume/world.py

@@ -166,6 +166,7 @@ async def setup_output_agent(self, simulation_id: str, save_frequency_hours: int
             db_engine=self.db,
             export_csv_path=self.export_csv_path,
             save_frequency_hours=save_frequency_hours,
+            learning_mode=self.learning_mode,
         )
         if self.same_process:
             output_agent = RoleAgent(

examples/inputs/example_01_rl/config.yml

@@ -41,3 +41,48 @@ base:
       minimum_bid_price: -500
       price_unit: EUR/MWh
       market_mechanism: pay_as_clear
+
+
+tiny:
+  start_date: 2019-01-01 00:00
+  end_date: 2019-01-02 00:00
+  time_step: 1h
+  save_frequency_hours: 24
+
+  learning_config:
+    observation_dimension: 50
+    action_dimension: 2
+    continue_learning: False
+    load_model_path: None
+    max_bid_price: 100
+    learning_mode: True
+    algorithm: matd3
+    learning_rate: 0.001
+    training_episodes: 2
+    collecting_initial_experience: 1
+    train_freq: 24
+    gradient_steps: -1
+    batch_size: 128
+    gamma: 0.99
+    device: cpu
+    noise_sigma: 0.1
+    noise_scale: 1
+    noise_dt: 1
+    validation_episodes_interval: 10
+
+  markets_config:
+    EOM:
+      operator: EOM_operator
+      product_type: energy
+      products:
+        - duration: 1h
+          count: 1
+          first_delivery: 1h
+      opening_frequency: 1h
+      opening_duration: 1h
+      volume_unit: MWh
+      maximum_bid_volume: 100000
+      maximum_bid_price: 3000
+      minimum_bid_price: -500
+      price_unit: EUR/MWh
+      market_mechanism: pay_as_clear

examples/inputs/example_01a/config.yml

@@ -20,6 +20,28 @@ base:
       price_unit: EUR/MWh
       market_mechanism: pay_as_clear
 
+tiny:
+  start_date: 2019-01-01 00:00
+  end_date: 2019-01-02 00:00
+  time_step: 1h
+  save_frequency_hours: 24
+  markets_config:
+    EOM:
+      operator: EOM_operator
+      product_type: energy
+      products:
+        - duration: 1h
+          count: 1
+          first_delivery: 1h
+      opening_frequency: 1h
+      opening_duration: 1h
+      volume_unit: MWh
+      maximum_bid_volume: 100000
+      maximum_bid_price: 3000
+      minimum_bid_price: -500
+      price_unit: EUR/MWh
+      market_mechanism: pay_as_clear
+
 dam:
   start_date: 2019-01-01 00:00
   end_date: 2019-04-01 00:00

pyproject.toml

@@ -1,8 +1,8 @@
 [tool.poetry]
-name = "assume_framework"
-version = "0.0.1"
-description = "ASSUME - Agent based energy market modeling using mango agents"
-authors = ["ASSUME developers <[email protected].de>"]
+name = "assume-framework"
+version = "0.1.0"
+description = "ASSUME - Agent-Based Electricity Markets Simulation Toolbox"
+authors = ["ASSUME Developers <contact@assume-project.de>"]
 license = "LICENSE"
 readme = "README.md"
 
@@ -14,10 +14,9 @@ keywords = ["agent based simulation", "energy market", "reinforcement learning",
 classifiers=[
     "Development Status :: 3 - Alpha",
     "Intended Audience :: Developers",
-    "License :: OSI Approved :: GNU Affero General Public License v3 (AGPLv3)",
+    "License :: OSI Approved :: GNU Affero General Public License v3 or later (AGPLv3+)",
     "Operating System :: OS Independent",
-    "Framework :: asyncio",
-    "Framework :: mango-agents",
+    "Framework :: AsyncIO",
     "Topic :: Scientific/Engineering",
     "Topic :: Software Development :: Libraries :: Python Modules",
 ]
@@ -29,7 +28,7 @@ packages = [
 [tool.poetry.dependencies]
 python = "^3.10"
 paho-mqtt = "^1.5.1"
-mango-agents = {git = "https://gitlab.com/maurerle/mango"}
+mango-agents = "^1.1.1"
 tqdm = "^4.64.1"
 python-dateutil = "^2.8.2"
 sqlalchemy = "^2.0.9"

tests/test_integration_cli.py

@@ -5,5 +5,12 @@
 
 @pytest.mark.slow
 def test_cli():
-    args = "-s example_01a -c base -db sqlite:///./examples/local_db/test_mini.db"
+    args = "-s example_01a -c tiny -db sqlite:///./examples/local_db/test_mini.db"
+    cli(args.split(" "))
+
+
+@pytest.mark.slow
+@pytest.mark.require_learning
+def test_cli_learning():
+    args = "-s example_01_rl -c tiny -db sqlite:///./examples/local_db/test_mini_rl.db"
     cli(args.split(" "))