Update README.md

ci-fixing/ci-fixing-benchmark/README.md

@@ -1,4 +1,4 @@
-Configs
+## Config
 
 To initialize the benchmark, you need to pass a path to a config file with the following fields (see example in benchmark.yaml):
 
@@ -7,7 +7,9 @@ To initialize the benchmark, you need to pass a path to a config file with the f
 **data_cache_dir**: here the cached dataset would be stored  
 **username**: your GitHub username  
 **test_username**: Username that would be displayed in the benchmark  
-**language**: # dataset language (now only Python is available)  
+**language**: dataset language (now only Python is available)  
+
+## Benchmark usage
 
 To use the benchmark you need to pass a function that fixes the repo according 
 the repo state and logs and metadata of the failed workflows (fix_repo_function).
@@ -20,20 +22,23 @@ It should have the following (all optional) arguments:
 **repo**:       git.Repo object from GitPython library  
 **out_folder**: folder for the benchmark results output  
 
-For now, I implemented only two functions:
+For now, only two functions have been implemented:
 
 **fix_none** -       does nothing  
-**fix_apply_diff** - applies the diff that fixed issue in the original repo  
+**fix_apply_diff** - applies the diff that fixed the issue in the original repo  
 
+## Evaluate dataset
 
-method CIFixBenchmark.eval_dataset(fix_repo_function) evaluates dataset:
+method **CIFixBenchmark.eval_dataset(fix_repo_function)** evaluates dataset:
 
 1. Downloads dataset (from https://huggingface.co/datasets/JetBrains-Research/lca-ci-fixing)
 2. Sends the datapoints on GitHub to run workflows
 3. Requests results from GitHub
 4. Analyzes results and print them.
 
-Further, maybe we will duplicate the request part on our side.
+Further, we may duplicate the request part on our side.
+
+## Outputs
 
 Method's outputs:
 
@@ -44,7 +49,7 @@ Method's outputs:
 
 Examples can be found in the examples folder /mnt/data/shared-data/lca/CI-fix-benchmark/benchmark/out
 
-You can also just evaluate your results using method CIFixBenchmark.eval_jobs(result_filename=result_filename)
+You can also evaluate your results using method **CIFixBenchmark.eval_jobs(result_filename=result_filename)**
 passing jobs_ids.jsonl file.
 
-You can download the dataset using CIFixBenchmark.get_dataset() method (example at the end of the file)
+You can download the dataset using **CIFixBenchmark.get_dataset()** method (example at the end of the file)