Skip to content

Commit 10f7312

Browse files
RashmiLnuRashmiLnu
authored
Feature: Add the cold run demo for the CommitHunter. (#1073)
* Feature: Add the cold run demo for the CommitHunter. Signed-off-by: Rashmi <Rashmi.Lnu@unh.edu> * Modify the cell * Move Version1.ipynb to CommitHunter folder as requested
1 parent 2adc316 commit 10f7312

1 file changed

Lines changed: 339 additions & 0 deletions

File tree

‎CommitHunter/Version1.ipynb‎

Lines changed: 339 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,339 @@
1+
{
2+
"cells": [
3+
{
4+
"cell_type": "markdown",
5+
"metadata": {
6+
"colab_type": "text",
7+
"id": "view-in-github"
8+
},
9+
"source": [
10+
"<a href=\"https://colab.research.google.com/github/RashmiLnu/aqa-test-tools/blob/Google_Collab_Rashmi/Version1.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
11+
]
12+
},
13+
{
14+
"cell_type": "code",
15+
"execution_count": null,
16+
"metadata": {
17+
"id": "NW7h6VD2xC6a"
18+
},
19+
"outputs": [],
20+
"source": [
21+
"!pip install -q langchain together openai faiss-cpu tiktoken\n"
22+
]
23+
},
24+
{
25+
"cell_type": "code",
26+
"execution_count": null,
27+
"metadata": {
28+
"id": "NtPGB4JfxQqt"
29+
},
30+
"outputs": [],
31+
"source": [
32+
"import os\n",
33+
"from langchain.llms import Together\n",
34+
"\n",
35+
"# Paste your Together.ai API key here (inside quotes)\n",
36+
"os.environ[\"TOGETHER_API_KEY\"] = \"Token_KEY\"\n",
37+
"\n",
38+
"# Load the LLM from Together.ai (you can switch models later)\n",
39+
"llm = Together(\n",
40+
" model=\"mistralai/Mistral-7B-Instruct-v0.1\",\n",
41+
" temperature=0.3,\n",
42+
" max_tokens=512\n",
43+
")\n"
44+
]
45+
},
46+
{
47+
"cell_type": "code",
48+
"execution_count": null,
49+
"metadata": {
50+
"id": "6e52bb9b"
51+
},
52+
"outputs": [],
53+
"source": [
54+
"!pip install -q langchain-community"
55+
]
56+
},
57+
{
58+
"cell_type": "code",
59+
"execution_count": null,
60+
"metadata": {
61+
"id": "f9e0ba93"
62+
},
63+
"outputs": [],
64+
"source": [
65+
"!pip install -U langchain-together"
66+
]
67+
},
68+
{
69+
"cell_type": "code",
70+
"execution_count": null,
71+
"metadata": {
72+
"id": "fe5a52e5"
73+
},
74+
"outputs": [],
75+
"source": [
76+
"import os\n",
77+
"from langchain_together import Together\n",
78+
"\n",
79+
"# Paste your Together.ai API key here (inside quotes)\n",
80+
"os.environ[\"TOGETHER_API_KEY\"] = \"Token\"\n",
81+
"\n",
82+
"# Load the LLM from Together.ai (you can switch models later)\n",
83+
"llm = Together(\n",
84+
" model=\"mistralai/Mistral-7B-Instruct-v0.1\",\n",
85+
" temperature=0.3,\n",
86+
" max_tokens=512\n",
87+
")"
88+
]
89+
},
90+
{
91+
"cell_type": "code",
92+
"execution_count": null,
93+
"metadata": {
94+
"id": "5W6agZt6yPYW"
95+
},
96+
"outputs": [],
97+
"source": [
98+
"from langchain.prompts import PromptTemplate\n",
99+
"from langchain.chains import LLMChain\n",
100+
"\n",
101+
"# Define prompt template\n",
102+
"template = \"\"\"\n",
103+
"You are a debugging assistant.\n",
104+
"Given the following test failure and commit diff, determine whether the commit likely caused the failure.\n",
105+
"\n",
106+
"Test Failure:\n",
107+
"{failure}\n",
108+
"\n",
109+
"Commit Diff:\n",
110+
"{diff}\n",
111+
"\n",
112+
"Answer:\n",
113+
"\"\"\"\n",
114+
"\n",
115+
"prompt = PromptTemplate(\n",
116+
" input_variables=[\"failure\", \"diff\"],\n",
117+
" template=template,\n",
118+
")\n",
119+
"\n",
120+
"test_failure = \"NullPointerException at NullHandler.java:42\"\n",
121+
"commit_diff = \"\"\"\n",
122+
"diff --git a/NullHandler.java b/NullHandler.java\n",
123+
"@@ -40,6 +40,7 @@\n",
124+
" public void handle() {\n",
125+
" Object obj = null;\n",
126+
" obj.toString(); // ← added line\n",
127+
" }\n",
128+
"\"\"\"\n",
129+
"\n",
130+
"\n",
131+
"chain = prompt | llm\n",
132+
"result = chain.invoke({\n",
133+
" \"failure\": test_failure,\n",
134+
" \"diff\": commit_diff\n",
135+
"})\n",
136+
"\n",
137+
"print(\"🤖 GPT Response:\\n\", result)\n"
138+
]
139+
},
140+
{
141+
"cell_type": "code",
142+
"execution_count": null,
143+
"metadata": {
144+
"id": "PnNaGqCW20Rt"
145+
},
146+
"outputs": [],
147+
"source": [
148+
"!pip install -q sentence-transformers\n"
149+
]
150+
},
151+
{
152+
"cell_type": "code",
153+
"execution_count": null,
154+
"metadata": {
155+
"id": "ni717d8U5e1C"
156+
},
157+
"outputs": [],
158+
"source": [
159+
"from sentence_transformers import SentenceTransformer\n",
160+
"import faiss\n",
161+
"import numpy as np\n",
162+
"\n",
163+
"# Load open-source embedding model\n",
164+
"embed_model = SentenceTransformer(\"all-MiniLM-L6-v2\") # or use bge-small-en\n",
165+
"\n",
166+
"# Simulate commit diffs\n",
167+
"commit_diffs = [\n",
168+
" (\"sha1\", \"Added null check to NullHandler.java\"),\n",
169+
" (\"sha2\", \"Refactored config loading\"),\n",
170+
" (\"sha3\", \"Removed unused function from TestRunner\"),\n",
171+
" (\"sha4\", \"Introduced obj.toString() in NullHandler.java\"),\n",
172+
" (\"sha5\", \"Added logging in exception handler\")\n",
173+
"]\n",
174+
"\n",
175+
"# Embed each diff\n",
176+
"texts = [text for _, text in commit_diffs]\n",
177+
"embeddings = embed_model.encode(texts, convert_to_numpy=True)\n",
178+
"\n",
179+
"# Build FAISS index\n",
180+
"dimension = embeddings.shape[1]\n",
181+
"index = faiss.IndexFlatL2(dimension)\n",
182+
"index.add(embeddings)\n",
183+
"\n",
184+
"# Store commit metadata separately\n",
185+
"sha_lookup = [sha for sha, _ in commit_diffs]\n"
186+
]
187+
},
188+
{
189+
"cell_type": "code",
190+
"execution_count": null,
191+
"metadata": {
192+
"id": "yQzf5L3l6HHN"
193+
},
194+
"outputs": [],
195+
"source": [
196+
"# Test failure text\n",
197+
"test_failure_context = \"NullPointerException at NullHandler.java:42 in handle()\"\n",
198+
"\n",
199+
"# Embed the failure\n",
200+
"query_embedding = embed_model.encode([test_failure_context])\n",
201+
"\n",
202+
"# Search for top 3 most similar diffs\n",
203+
"D, I = index.search(query_embedding, 3)\n",
204+
"top_matches = [(sha_lookup[i], commit_diffs[i][1]) for i in I[0]]\n",
205+
"\n",
206+
"print(\"🔍 Top relevant commit diffs:\\n\")\n",
207+
"for sha, diff in top_matches:\n",
208+
" print(f\"Commit {sha}:\\n{diff}\\n\")\n"
209+
]
210+
},
211+
{
212+
"cell_type": "code",
213+
"execution_count": null,
214+
"metadata": {
215+
"id": "0LiPUzbA7-WK"
216+
},
217+
"outputs": [],
218+
"source": [
219+
"from langchain.prompts import PromptTemplate\n",
220+
"\n",
221+
"template = \"\"\"\n",
222+
"You are a debugging assistant.\n",
223+
"\n",
224+
"A test failed with the following message:\n",
225+
"{failure}\n",
226+
"\n",
227+
"Here is a commit diff that was semantically similar:\n",
228+
"{diff}\n",
229+
"\n",
230+
"Do you think this commit caused the failure? Why or why not?\n",
231+
"\"\"\"\n",
232+
"\n",
233+
"prompt = PromptTemplate(\n",
234+
" input_variables=[\"failure\", \"diff\"],\n",
235+
" template=template,\n",
236+
")\n",
237+
"\n",
238+
"chain = prompt | llm\n"
239+
]
240+
},
241+
{
242+
"cell_type": "code",
243+
"execution_count": null,
244+
"metadata": {
245+
"id": "I7C-duv58Bdx"
246+
},
247+
"outputs": [],
248+
"source": [
249+
"print(\"🤖 GPT-style analysis:\\n\")\n",
250+
"\n",
251+
"for sha, diff in top_matches:\n",
252+
" print(f\"🧩 Commit {sha}:\")\n",
253+
" response = chain.invoke({\n",
254+
" \"failure\": test_failure_context,\n",
255+
" \"diff\": diff\n",
256+
" })\n",
257+
" print(response)\n",
258+
" print(\"-\" * 60)\n"
259+
]
260+
},
261+
{
262+
"cell_type": "code",
263+
"execution_count": null,
264+
"metadata": {
265+
"id": "fGZLq7VdARod"
266+
},
267+
"outputs": [],
268+
"source": [
269+
"combined_diff_text = \"\"\n",
270+
"\n",
271+
"for sha, diff in top_matches:\n",
272+
" combined_diff_text += f\"--- Commit {sha} ---\\n{diff}\\n\\n\"\n"
273+
]
274+
},
275+
{
276+
"cell_type": "code",
277+
"execution_count": null,
278+
"metadata": {
279+
"id": "fWmt6Ii0ATHp"
280+
},
281+
"outputs": [],
282+
"source": [
283+
"from langchain.prompts import PromptTemplate\n",
284+
"\n",
285+
"combined_template = \"\"\"\n",
286+
"You are a debugging assistant helping identify the root cause of a test failure.\n",
287+
"\n",
288+
"Here is the test failure:\n",
289+
"{failure}\n",
290+
"\n",
291+
"Here are the top 3 semantically related commits:\n",
292+
"{diffs}\n",
293+
"\n",
294+
"Based on the diff content and the test failure, which commit is most likely responsible? Explain your reasoning.\n",
295+
"\"\"\"\n",
296+
"\n",
297+
"combined_prompt = PromptTemplate(\n",
298+
" input_variables=[\"failure\", \"diffs\"],\n",
299+
" template=combined_template\n",
300+
")\n",
301+
"\n",
302+
"# Create runnable chain\n",
303+
"combined_chain = combined_prompt | llm\n"
304+
]
305+
},
306+
{
307+
"cell_type": "code",
308+
"execution_count": null,
309+
"metadata": {
310+
"id": "eQ6iKc4iAWSv"
311+
},
312+
"outputs": [],
313+
"source": [
314+
"result = combined_chain.invoke({\n",
315+
" \"failure\": test_failure_context,\n",
316+
" \"diffs\": combined_diff_text\n",
317+
"})\n",
318+
"\n",
319+
"print(\"🤖 Final Verdict:\\n\")\n",
320+
"print(result)\n"
321+
]
322+
}
323+
],
324+
"metadata": {
325+
"colab": {
326+
"include_colab_link": true,
327+
"provenance": []
328+
},
329+
"kernelspec": {
330+
"display_name": "Python 3",
331+
"name": "python3"
332+
},
333+
"language_info": {
334+
"name": "python"
335+
}
336+
},
337+
"nbformat": 4,
338+
"nbformat_minor": 0
339+
}

0 commit comments

Comments
 (0)