Merge development of Example 1 (and various train/infer/eval scripts) into Master

Major changes:

• Large code additions to each stage, these may not be compatible yet with the version that Sylvain will write, but should be able to merge in quite easily
• train.py, infer.py, eval.py complete as version 1. These may also need to change after Sylvain overall's class updates
• examples/Example_1 now has a set of .yaml files for each stage's train/infer/eval process

examples/Example_1/README.md

+# Example 1
+
+## Setup
+
+Assuming the CommonFramework repo requirements have been installed, the only other requirements are the data files and model files. An example dataset can be downloaded to a data directory `MY_DATA_DIR`. This is around 2.3Gb, so ensure you download to a location with sufficient space. Define this dir with
+```bash
+data_dir=MY_DATA_DIR
+```
+then download the data with
+```bash
+mkdir $data_dir/Example_1
+wget https://cernbox.cern.ch/remote.php/dav/public-files/AREZqMSHGrWMIjc/athena_100_events.zip -O $data_dir/Example_1/athena_100_events.zip
+unzip $data_dir/Example_1/athena_100_events.zip -d $data_dir/Example_1
+```
+and enter the password provided by the GNN4ITk team. The model files can be downloaded with
+```bash
+wget https://cernbox.cern.ch/remote.php/dav/public-files/uUCHDnGUiHdhsyl/Example_1.zip -O $data_dir/Example_1/Example_1.zip
+unzip $data_dir/Example_1/Example_1.zip -d $data_dir/Example_1
+```
+and enter the password provided by the GNN4ITk team. The location of this data, as well as all parameters controlling the GNN4ITk reconstruction chain, is specified in `yaml` config files. The data directory currently has a placeholder MY_DATA_DIR. Replace this with the actual data directory with
+```bash
+sed -i "s/MY_DATA_DIR/$data_dir/g" *.yaml
+```
+
+## Running the Example
+
+The following commands will run the Example 1 pipeline. In general, they follow the pattern
+```
+train --> infer --> eval
+``` 
+where `train` is used to train a model, `infer` is used to apply the model to data, and `eval` is used to evaluate the performance of the model. If a model has already been trained (in the case of the Module Map in Example 1), we do not need to train it, only provide the model to the `infer` step.
+
+**1.** First, we build our input data from the raw Athena events:
+```bash
+g4i-infer data_reader.yaml
+```
+
+**2.** Then, we build graphs using the python implementation of the Module Map:
+```bash
+g4i-infer module_map_infer.yaml
+```
+
+**2a.** (Optional) We can examine the performance of the Module Map, by printing some efficiency plots:
+```bash
+g4i-eval module_map_eval.yaml
+```
+
+**3.** Then, we train the GNN (here we will train on a toy version of the data that only includes pT>1GeV particles - this is configured with the `hard_cuts` option in the `gnn_train.yaml` file):
+```bash
+g4i-train gnn_train.yaml
+```
+
+**4.** Once the GNN is trained (should take around half an hour), we apply the GNN in inference to produce a dataset of scored graphs:
+```bash
+g4i-infer gnn_infer.yaml
+```
+
+**5.** Finally, we produce track candidates from the scored graphs:
+```bash
+g4i-infer track_building_infer.yaml
+```
+
+**6.** And plot the performance
+```bash
+g4i-eval track_building_eval.yaml
+```
+
+## Understanding the Example
+