Major revamp - new grey box metrics (#33)

Added new grey box metrics for evaluation for RAG without ground truth.

Author: shahules786

.gitattributes

@@ -0,0 +1 @@
+*.tar.gz filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -158,4 +158,9 @@ cython_debug/
 #  and can be added to the global gitignore or merged into this file.  For a more nuclear
 #  option (not recommended) you can uncomment the following to ignore the entire idea folder.
 #.idea/
+
+# Ragas specific
 ragas/_version.py
+experiments/**/data
+experiments/**/storage
+**/fil-result/

experiments/README.md

@@ -0,0 +1,12 @@
+# Experiments
+:test_tube:	**This section is for experiments**
+
+We commit and record results from different experiments done using Ragas here. We do two types of experiments,
+1. `assesments` : To develop better metrics,paradigms and evaluate metrics against human judgment.
+2. `baselines`: Create RAG pipelines and evaluate them using Ragas.
+
+Any machine learning comprehension dataset can be used for experiments. Add any newly tried dataset to below list.
+Datasets:
+1. [HotpotQA](https://huggingface.co/datasets/hotpot_qa)
+2. [WikiQA test](explodinggradients/ragas-wikiqa)
+