FEATURE: Adding the first commit of main

.github/workflows/huggingface.yaml

@@ -0,0 +1,17 @@
+name: Sync to Hugging Face hub
+on:
+  push:
+    branches: [main]
+
+jobs:
+  sync-to-hub:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+          lfs: true
+      - name: Push to hub
+        env:
+          HF_TOKEN: ${{ secrets.HF_TOKEN }}
+        run: git push --force https://${{ secrets.HF_USERNAME }}:[email protected]/spaces/${{ secrets.HF_USERNAME }}/${{ secrets.SPACE_NAME }} main

.gitignore

@@ -0,0 +1,58 @@
+# Archivos de Python
+__pycache__/
+*.py[cod]
+*.pyo
+*.pyd
+*.pdb
+*.egg-info/
+*.egg
+*.manifest
+*.spec
+
+# Entornos virtuales
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Archivos de configuración/editor
+.vscode/
+.idea/
+*.swp
+*.swo
+
+# Archivos de compilación
+build/
+dist/
+.eggs/
+*.so
+
+# Logs y cachés
+*.log
+*.pot
+*.pyc
+*.tmp
+*.DS_Store
+.cache/
+coverage/
+htmlcov/
+.tox/
+.mypy_cache/
+.pytest_cache/
+
+# Jupyter Notebooks
+.ipynb_checkpoints
+*.ipynb_checkpoints
+
+# Otros
+*.sqlite3
+*.db
+*.mo
+
+*.parquet
+*.xlsx
+*.json
+**/AutogluonModels/**

.pre-commit-config.yaml

@@ -0,0 +1,10 @@
+repos:
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    rev: v0.1.15
+    hooks:
+      - id: ruff
+        types_or: [python, pyi, jupyter]
+        args: 
+          - --fix
+      - id: ruff-format
+        types_or: [python, pyi, jupyter]

Dockerfile

@@ -0,0 +1,17 @@
+FROM python:3.11-slim
+
+WORKDIR /app
+COPY requirements.txt requirements.txt
+COPY pyproject.toml pyproject.toml
+COPY src/ src/
+COPY data/ /data
+
+RUN pip install uv
+RUN uv pip install -r requirements.txt
+RUN pip install -e .
+
+EXPOSE 8501
+
+HEALTHCHECK CMD curl --fail http://localhost:8501/_stcore/health
+
+ENTRYPOINT ["streamlit", "run", "src/streamlit_app.py", "--server.port=8501", "--server.address=0.0.0.0"]

README.md

@@ -0,0 +1,39 @@
+---
+title: Amazon Chronos Bolt
+emoji: 👁
+colorFrom: gray
+colorTo: purple
+sdk: docker
+pinned: false
+license: mit
+short_description: Large Time-Series Models Amazon 
+---
+
+# Modelos Fundacionales para Series de tiempo
+
+Conferencia: **Aprendiendo el Lenguaje de las Series de Tiempo: AWS Chronos Bolt.**
+
+## Descripción
+
+Este repo contiene el benchmark y código de la aplicación con datos del INAMHI de la conferencia `Aprendiendo el Lenguaje de las Series de Tiempo: AWS Chronos Bolt.` dada durante el AWS Community Day Ecuador 2025. 
+
+AWS Chronos Bolt es un modelo open-source. Cualquier persona interesada, puede acceder en los siguientes links a las versiones disponibles del modelo.
+
+- [Tiny](https://huggingface.co/amazon/chronos-bolt-tiny)
+- [Mini](https://huggingface.co/amazon/chronos-bolt-mini)
+- [Small](https://huggingface.co/amazon/chronos-bolt-small)
+- [Base](https://huggingface.co/amazon/chronos-bolt-base)
+
+## ☕ Support the Project
+
+If you find this demo helpful and would like to support its continued development, consider buying me a coffee. Your support helps maintain and improve this project!
+
+[![Buy Me A Coffee](https://www.buymeacoffee.com/assets/img/custom_images/orange_img.png)](https://www.paypal.com/paypalme/sebassarasti)
+
+### Other Ways to Support
+- ⭐ Star this repository
+- 🍴 Fork it and contribute
+- 📢 Share it with others who might find it useful
+- 🐛 Report issues or suggest new features
+
+Your support, in any form, is greatly appreciated! 🙏

pyproject.toml

@@ -0,0 +1,31 @@
+[project]
+name = "chronos_conference"
+version = "0.1.0"
+description = "A Python package to execute most common models for retail"
+authors = [
+  { name = "Sebastian Sarasti", email = "[email protected]" }
+]
+readme = "README.md"
+
+requires-python = ">=3.11"
+dependencies = [
+    "autogluon==1.4.0",
+    "kagglehub==0.3.13",
+    "numpy==2.1.3",
+    "pandas==2.3.2",
+    "utilsforecast==0.2.11",
+    "statsforecast==2.0.1",
+    "ipywidgets",
+    "ipykernel",
+    "pre_commit==4.3.0",
+    "streamlit==1.50.0",
+    "plotly==6.3.0",
+    "nbformat==5.10.4"
+]
+
+[project.urls]
+Repository = "https://github.com/sebassaras02/AWS_Community_Day_Ecuador_2025"
+
+[build-system]
+requires = ["setuptools>=61.0", "wheel"]
+build-backend = "setuptools.build_meta"

requirements.txt

@@ -0,0 +1,12 @@
+autogluon==1.4.0
+kagglehub==0.3.13
+numpy==2.1.3
+pandas==2.3.2
+utilsforecast==0.2.11
+statsforecast==2.0.1
+ipywidgets
+ipykernel
+pre_commit==4.3.0
+streamlit==1.50.0
+plotly==6.3.0
+nbformat==5.10.4

src/chronos_conference/adapters/filter_ts.py

@@ -0,0 +1,10 @@
+import pandas as pd
+
+
+def filter_ts(
+    df: pd.DataFrame, date_col: str, min_date: str, max_date: str
+) -> pd.DataFrame:
+    df = df.copy()
+    if pd.api.types.is_datetime64_any_dtype(df[date_col]) is False:
+        df[date_col] = pd.to_datetime(df[date_col])
+    return df[(df[date_col] >= min_date) & (df[date_col] <= max_date)]

src/chronos_conference/adapters/model_instance.py

@@ -0,0 +1,52 @@
+from abc import ABC
+
+import pandas as pd
+from autogluon.timeseries import TimeSeriesDataFrame, TimeSeriesPredictor
+
+
+class ForecastingBaseModel(ABC):
+    def __init__(self, freq: str, n_jobs: int = 1) -> None:
+        self.model = None
+        self.freq = freq
+        self.n_jobs = n_jobs
+
+    def fit(
+        self, df: pd.DataFrame, date_col: str, item_col: str, targe_col: str
+    ) -> None:
+        pass
+
+    def predict(self, n_steps: int) -> pd.DataFrame:
+        pass
+
+
+class ChronosForecaster(ForecastingBaseModel):
+    def __init__(self, freq: str = "H"):
+        super().__init__(freq=freq)
+
+    def fit(
+        self, df: pd.DataFrame, date_col: str, item_col: str, target_col: str
+    ) -> None:
+        self.item_id = item_col
+        df = df.copy()
+        df = df.rename(columns={target_col: "target"})
+        self.df = TimeSeriesDataFrame.from_data_frame(
+            df,
+            id_column=item_col,
+            timestamp_column=date_col,
+        )
+
+    def predict(self, n_steps):
+        self.model = TimeSeriesPredictor(
+            prediction_length=n_steps, freq=self.freq, verbosity=0
+        ).fit(self.df, presets="bolt_base")
+        results = self.model.predict(self.df)
+        results = results.to_data_frame().reset_index()
+        results = results[["mean", "item_id", "timestamp"]]
+        results = results.rename(
+            columns={
+                "mean": "AWSChronosForecast",
+                "item_id": "unique_id",
+                "timestamp": "ds",
+            }
+        )
+        return results

src/chronos_conference/adapters/ts_plot.py

@@ -0,0 +1,16 @@
+import plotly.express as px
+import pandas as pd
+
+
+def get_plot(df_story: pd.DataFrame, df_pred: pd.DataFrame):
+    if pd.api.types.is_datetime64_any_dtype(df_story["datetime"]) is False:
+        df_story["datetime"] = pd.to_datetime(df_story["datetime"])
+
+    if pd.api.types.is_datetime64_any_dtype(df_pred["ds"]) is False:
+        df_pred["ds"] = pd.to_datetime(df_pred["ds"])
+
+    fig = px.line(df_story, x="datetime", y="value", title="Historical Information")
+    fig.add_scatter(
+        x=df_pred["ds"], y=df_pred["AWSChronosForecast"], mode="lines", name="Forecast"
+    )
+    return fig

src/chronos_conference/domain/inference.py

@@ -0,0 +1,25 @@
+import pandas as pd
+
+
+def get_forecast(
+    df: pd.DataFrame,
+    date_col: str,
+    target_col: str,
+    item_col: str,
+    model_instance,
+    horizon: int = 24,
+) -> pd.DataFrame:
+    """
+    This function is designed to generate forecasts using a pre-trained
+    TimeSeriesPredictor model from the AutoGluon library.
+    """
+    df = df.copy()
+    df = df.rename(columns={target_col: "target"})
+    model_instance.fit(
+        df=df,
+        date_col=date_col,
+        item_col=item_col,
+        target_col=target_col,
+    )
+    results = model_instance.predict(n_steps=horizon)
+    return results

src/chronos_conference/service_layer/main.py

@@ -0,0 +1,73 @@
+# ruff: noqa: F403, F405
+
+import streamlit as st
+import pandas as pd
+
+from chronos_conference.domain.inference import get_forecast
+from chronos_conference.adapters.filter_ts import filter_ts
+from chronos_conference.adapters.model_instance import ChronosForecaster
+from chronos_conference.adapters.ts_plot import get_plot
+from chronos_conference.settings import *
+
+st.title("AWS Community Day Ecuador 2025")
+st.header(
+    "Conferencia: Aprendiendo el Lenguaje de las series de tiempo con AWS Chronos Bolt"
+)
+st.subheader("Ponente: Sebastian Sarasti")
+
+st.write(
+    """
+Esta aplicación demuestra cómo utilizar AWS Chronos Bolt para la predicción del clima mediante
+datos abiertos obtenidos del INAMHI. 
+"""
+)
+
+df = pd.read_csv("../../../data/historical_simulation_9023624.csv")
+
+col1, col2, col3 = st.columns(3)
+
+with col1:
+    min_date = st.date_input("Fecha mínima", value=MIN_PRED_DATE)
+
+with col2:
+    max_date = st.date_input("Fecha máxima", value=MAX_PRED_DATE)
+
+with col3:
+    n_steps = st.number_input(
+        "Número de pasos a predecir",
+        min_value=MIN_PRED_DATE_LIMIT,
+        max_value=MAX_PRED_DATE_LIMIT,
+        value=N_PRED_STEPS,
+    )
+
+execution_button = st.button("Ejecutar modelo")
+
+if not execution_button:
+    st.stop()
+
+with st.spinner("Filtrando datos..."):
+    df_useful = filter_ts(
+        df,
+        date_col=HISTORICAL_DATE_COLUMN,
+        min_date=str(min_date),
+        max_date=str(max_date),
+    )
+
+model = ChronosForecaster(freq=FREQUENCY)
+
+with st.spinner("Modelo en ejecución..."):
+    results = get_forecast(
+        df=df_useful,
+        date_col=HISTORICAL_DATE_COLUMN,
+        target_col=HISTORICAL_TARGET_COLUMN,
+        item_col=HISTORICAL_ITEM_COLUMN,
+        model_instance=model,
+    )
+
+st.success("¡Ejecución completada!")
+
+st.write("Resultados de la ejecución:")
+
+fig = get_plot(df_useful, results)
+
+st.plotly_chart(fig, use_container_width=True)

src/chronos_conference/settings.py

@@ -0,0 +1,11 @@
+HISTORICAL_DATE_COLUMN = "datetime"
+HISTORICAL_ITEM_COLUMN = "unique_id"
+HISTORICAL_TARGET_COLUMN = "value"
+
+FREQUENCY = "D"
+
+MIN_PRED_DATE = "2022-01-01"
+MAX_PRED_DATE = "2023-06-30"
+N_PRED_STEPS = 48
+MIN_PRED_DATE_LIMIT = 1
+MAX_PRED_DATE_LIMIT = 128