extend-ben-dataset branch changes

src/python/diff3_analysis.ipynb

@@ -0,0 +1,70 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "dropwizard/metrics  : Cloning repo\n",
+      "dropwizard/metrics  : Finished cloning\n",
+      "dropwizard/metrics  : Finished cloning\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Switched to branch 'TEMP_LEFT_BRANCH'\n",
+      "diff: ./repos/merge_attempt/dropwizard/metrics/pom.xml: No such file or directory\n",
+      "diff: ./repos/programmer_merge/dropwizard/metrics/pom.xml: No such file or directory\n"
+     ]
+    }
+   ],
+   "source": [
+    "repo_num = 548\n",
+    "merge_tool = \"gitmerge_ort\"\n",
+    "# merge_tool = \"gitmerge_ort_adjacent\"\n",
+    "# merge_tool = \"gitmerge_ort_ignorespace\"\n",
+    "# merge_tool = \"gitmerge_ort_imports\"\n",
+    "# merge_tool = \"gitmerge_ort_imports_ignorespace\"\n",
+    "# merge_tool = \"gitmerge_resolve\"\n",
+    "# merge_tool = \"gitmerge_recursive_histogram\"\n",
+    "# merge_tool = \"gitmerge_recursive_ignorespace\"\n",
+    "# merge_tool = \"gitmerge_recursive_minimal\"\t\n",
+    "# merge_tool = \"gitmerge_recursive_myers\"\n",
+    "# merge_tool = \"gitmerge_recursive_patience\"\n",
+    "# merge_tool = \"git_hires_merge\"\n",
+    "# merge_tool = \"spork\"\n",
+    "# merge_tool = \"intellimerge\"\n",
+    "\n",
+    "from diff3_analysis import diff3_analysis\n",
+    "diff3_analysis(merge_tool, repo_num)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "research",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.8.18"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

src/python/diff3_analysis.py

@@ -0,0 +1,139 @@
+"""Runs a merge and uses diff3 to compare it to the base and final branch of a given repo.
+"""
+import subprocess
+import re
+import os
+import tempfile
+import pandas as pd
+from repo import clone_repo_to_path
+from merge_tester import MERGE_STATE
+
+# pylint: disable-msg=too-many-locals
+
+
+def diff3_analysis(merge_tool: str, repo_num: int):
+    """
+    Analyzes merge conflicts using the diff3 tool and opens the results in the default text viewer.
+
+    Args:
+        merge_tool (str): The merge tool to be used.
+        repo_num (int): The index of the repository in the results DataFrame.
+
+    Returns:
+        None
+    """
+    df = pd.read_csv("../../results_greatest_hits/result.csv")
+    repo_name = df.iloc[repo_num]["repository"]
+
+    script = "../scripts/merge_tools/" + merge_tool + ".sh"
+    repo = clone_repo_to_path(
+        repo_name, "./repos/merge_attempt"
+    )  # Return a Git-Python repo object
+    repo.remote().fetch()
+    left_sha = df.iloc[repo_num]["left"]
+    repo.git.checkout(left_sha, force=True)
+    repo.submodule_update()
+    repo.git.checkout("-b", "TEMP_LEFT_BRANCH", force=True)
+    repo.git.checkout(df.iloc[repo_num]["right"], force=True)
+    repo.submodule_update()
+    repo.git.checkout("-b", "TEMP_RIGHT_BRANCH", force=True)
+
+    result = subprocess.run(
+        [
+            script,
+            repo.git.rev_parse("--show-toplevel"),
+            "TEMP_LEFT_BRANCH",
+            "TEMP_RIGHT_BRANCH",
+        ],
+        stdout=subprocess.PIPE,
+        text=True,
+    )
+
+    conflict_file_matches = re.findall(
+        r"CONFLICT \(.+\): Merge conflict in (.+)", result.stdout
+    )
+
+    repo = clone_repo_to_path(
+        repo_name, "./repos/programmer_merge"
+    )  # Return a Git-Python repo object
+    repo.git.checkout(df.iloc[repo_num]["merge"], force=True)
+    repo.submodule_update()
+
+    '''
+    repo = clone_repo_to_path(
+        repo_name, "./repos/base"
+    )  # Return a Git-Python repo object
+    repo.git.checkout(df.iloc[repo_num]["base"], force=True)
+    repo.submodule_update()
+    '''
+
+    for conflict_file_match in conflict_file_matches:
+        conflicting_file = str(conflict_file_match)
+        conflict_path = os.path.join(repo_name, conflicting_file)
+        conflict_path_merge_attempt = os.path.join(
+            "./repos/merge_attempt", conflict_path
+        )
+
+        '''
+        conflict_path_base = os.path.join("./repos/base", conflict_path)
+        '''
+        conflict_path_programmer_merge = os.path.join(
+            "./repos/programmer_merge", conflict_path
+        )
+        '''
+        diff_results = subprocess.run(
+            [
+                "diff3",
+                conflict_path_base,
+                conflict_path_merge_attempt,
+                conflict_path_programmer_merge,
+            ],
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            text=True,
+        )
+
+        # Check that diff3 didn't run into missing files in the base
+        error_message = "No such file or directory"
+        if error_message in diff_results.stderr:
+        '''
+        # Since the conflict file was added in both parents we can't diff the base.
+        diff_results = subprocess.run(
+            [
+                "diff",
+                conflict_path_merge_attempt,
+                conflict_path_programmer_merge,
+            ],
+            stdout=subprocess.PIPE,
+            text=True,
+        )
+
+        # Use a temporary file to store the diff results
+        with tempfile.NamedTemporaryFile(mode="w+", delete=False) as temp_file:
+            temp_file.write(diff_results.stdout)
+
+        # Open the saved text file with the default application
+        subprocess.run(["xdg-open", temp_file.name], check=True)
+
+        # Delete the temporary file
+        os.remove(temp_file.name)
+
+    # Deletes base, programmer_merge, and merge_attempt folders in repos dir
+    # We do this to prevent errors if cloning the same repo into the folder twice
+    '''
+    subprocess.run(
+        ["rm", "-rf", "./repos/base"],
+        stderr=subprocess.PIPE,
+        stdout=subprocess.PIPE,
+    )
+    '''
+    subprocess.run(
+        ["rm", "-rf", "./repos/merge_attempt"],
+        stderr=subprocess.PIPE,
+        stdout=subprocess.PIPE,
+    )
+    subprocess.run(
+        ["rm", "-rf", "./repos/programmer_merge"],
+        stderr=subprocess.PIPE,
+        stdout=subprocess.PIPE,
+    )

src/python/repo.py

@@ -57,6 +57,34 @@ def clone_repo(repo_slug: str) -> git.repo.Repo:
     return repo
 
 
+def clone_repo_to_path(repo_slug: str, path: str) -> git.repo.Repo:
+    """Clones a repository, or runs `git fetch` if the repository is already cloned.
+    Args:
+        repo_slug (str): The slug of the repository, which is "owner/reponame".
+    """
+    repo_dir = REPOS_PATH / Path(repo_slug)
+    if repo_dir.exists():
+        repo = git.repo.Repo(repo_dir)
+    else:
+        repo_dir.parent.mkdir(parents=True, exist_ok=True)
+        os.environ["GIT_TERMINAL_PROMPT"] = "0"
+        print(repo_slug, " : Cloning repo")
+        # ":@" in URL ensures that we are not prompted for login details
+        # for the repos that are now private.
+        github_url = "https://:@github.com/" + repo_slug + ".git"
+        print(repo_slug, " : Finished cloning")
+        try:
+            repo = git.repo.Repo.clone_from(github_url, repo_dir)
+            print(repo_slug, " : Finished cloning")
+            repo.remote().fetch()
+            repo.remote().fetch("refs/pull/*/head:refs/remotes/origin/pull/*")
+            repo.submodule_update()
+        except Exception as e:
+            print(repo_slug, "Exception during cloning:\n", e)
+            raise
+    return repo
+
+
 TEST_STATE = Enum(
     "TEST_STATE",
     [