my first commit

Dockerfile

@@ -0,0 +1,46 @@
+# Build from a LINUX lightweight version of Anaconda
+FROM continuumio/miniconda3
+
+# Update packages and install nano unzip and curl
+RUN apt-get update
+RUN apt-get install nano unzip curl -y
+
+# Install AWS cli - Necessary since we are going to interact with S3 
+RUN curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"
+RUN unzip awscliv2.zip
+RUN ./aws/install
+
+# THIS IS SPECIFIC TO HUGGINFACE
+# We create a new user named "user" with ID of 1000
+RUN useradd -m -u 1000 user
+# We switch from "root" (default user when creating an image) to "user" 
+USER user
+# We set two environmnet variables 
+# so that we can give ownership to all files in there afterwards
+# we also add /home/user/.local/bin in the $PATH environment variable 
+# PATH environment variable sets paths to look for installed binaries
+# We update it so that Linux knows where to look for binaries if we were to install them with "user".
+ENV HOME=/home/user \
+    PATH=/home/user/.local/bin:$PATH
+
+# We set working directory to $HOME/app (<=> /home/user/app)
+WORKDIR $HOME/app
+
+# Copy and install dependencies 
+COPY requirements.txt /requirements.txt
+RUN pip install -r /requirements.txt
+
+# Copy all local files to /home/user/app with "user" as owner of these files
+# Always use --chown=user when using HUGGINGFACE to avoid permission errors
+COPY --chown=user . $HOME/app
+
+
+# Launch mlflow server 
+# Here we chose to have $PORT as environment variable but you could have hard coded 7860 
+# If you are sure to push into production 
+# Advantage to use an env variable is that your code is more portable if you were to deploy to another 
+# type of server
+CMD mlflow server -p $PORT \
+    --host 0.0.0.0 \
+    --backend-store-uri $BACKEND_STORE_URI \
+    --default-artifact-root $ARTIFACT_STORE_URI

requirements.txt

@@ -0,0 +1,10 @@
+boto3
+pandas 
+gunicorn 
+streamlit 
+scikit-learn 
+matplotlib 
+seaborn 
+plotly
+mlflow
+psycopg2-binary