From 63824779e39cd7ee175ba71f328fbdf5130892bc Mon Sep 17 00:00:00 2001
From: Ivar Holmlund <89637692+ivholmlu@users.noreply.github.com>
Date: Thu, 27 Jun 2024 00:02:02 +0200
Subject: [PATCH] Added utility for saving results to specified folder (#14375)

---
 .../llama_index/packs/rag_evaluator/base.py    | 18 ++++++++++++++++--
 .../pyproject.toml                             |  2 +-
 2 files changed, 17 insertions(+), 3 deletions(-)

diff --git a/llama-index-packs/llama-index-packs-rag-evaluator/llama_index/packs/rag_evaluator/base.py b/llama-index-packs/llama-index-packs-rag-evaluator/llama_index/packs/rag_evaluator/base.py
index 0fc5c7ee68ba7..973baaab43b7b 100644
--- a/llama-index-packs/llama-index-packs-rag-evaluator/llama_index/packs/rag_evaluator/base.py
+++ b/llama-index-packs/llama-index-packs-rag-evaluator/llama_index/packs/rag_evaluator/base.py
@@ -4,6 +4,8 @@
 import warnings
 from collections import deque
 from typing import Any, List, Optional
+import os
+from pathlib import Path
 
 import pandas as pd
 import tqdm
@@ -44,6 +46,7 @@ def __init__(
         judge_llm: Optional[LLM] = None,
         embed_model: Optional[BaseEmbedding] = None,
         show_progress: bool = True,
+        result_path: Optional[str] = None,
     ):
         self.query_engine = query_engine
         self.rag_dataset = rag_dataset
@@ -64,6 +67,15 @@ def __init__(
         }
         self.eval_queue = deque(range(len(rag_dataset.examples)))
         self.prediction_dataset = None
+        if result_path is None:
+            self.result_path = Path.cwd()
+        else:
+            self.result_path = Path(result_path)
+            if not self.result_path.is_absolute():
+                self.result_path = Path.cwd() / self.result_path
+
+        if not os.path.exists(self.result_path):
+            os.makedirs(self.result_path)
 
     async def _amake_predictions(
         self,
@@ -223,7 +235,9 @@ def _save_evaluations(self):
             "relevancy": [e.dict() for e in self.evals["relevancy"]],
         }
 
-        with open("_evaluations.json", "w") as json_file:
+        with open(
+            os.path.join(self.result_path, "_evaluations.json"), "w"
+        ) as json_file:
             json.dump(evaluations_objects, json_file)
 
     def _prepare_and_save_benchmark_results(self):
@@ -263,7 +277,7 @@ def _prepare_and_save_benchmark_results(self):
         mean_scores_df.index = mean_scores_df.index.set_names(["metrics"])
 
         # save mean_scores_df
-        mean_scores_df.to_csv("benchmark.csv")
+        mean_scores_df.to_csv(os.path.join(self.result_path, "benchmark.csv"))
         return mean_scores_df
 
     def _make_evaluations(
diff --git a/llama-index-packs/llama-index-packs-rag-evaluator/pyproject.toml b/llama-index-packs/llama-index-packs-rag-evaluator/pyproject.toml
index b91ae04d35261..55a8a92cd82a0 100644
--- a/llama-index-packs/llama-index-packs-rag-evaluator/pyproject.toml
+++ b/llama-index-packs/llama-index-packs-rag-evaluator/pyproject.toml
@@ -29,7 +29,7 @@ license = "MIT"
 maintainers = ["nerdai"]
 name = "llama-index-packs-rag-evaluator"
 readme = "README.md"
-version = "0.1.5"
+version = "0.1.6"
 
 [tool.poetry.dependencies]
 python = ">=3.8.1,<4.0"