File size: 5,979 Bytes
3d968cd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Synthetic Sensitive Data in Source Code (N=300) — Starter\n",
    "\n",
    "Quick exploration of the dataset: load samples, inspect categories, and preview labeled secrets.\n",
    "\n",
    "**Intended use:** secret/PII detection, local masking evaluation, OWASP LLM02–style leakage tests.\n",
    "\n",
    "> All values are **synthetic**. Do not treat them as real credentials."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pathlib import Path\n",
    "import json\n",
    "import pandas as pd\n",
    "\n",
    "# Works on Kaggle and locally\n",
    "CANDIDATES = [\n",
    "    Path(\"/kaggle/input\"),\n",
    "    Path(\".\"),\n",
    "]\n",
    "\n",
    "csv_path = None\n",
    "json_path = None\n",
    "for root in CANDIDATES:\n",
    "    hits = list(root.rglob(\"synthetic_sensitive_data_in_source_code_n300.csv\"))\n",
    "    if hits:\n",
    "        csv_path = hits[0]\n",
    "        json_path = csv_path.with_suffix(\".json\")\n",
    "        break\n",
    "\n",
    "assert csv_path is not None, \"Dataset CSV not found\"\n",
    "print(\"CSV :\", csv_path)\n",
    "print(\"JSON:\", json_path if json_path.exists() else \"(optional)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Load CSV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(csv_path)\n",
    "print(df.shape)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Category & language distribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(\"Categories\")\n",
    "display(df[\"category\"].value_counts().rename(\"count\").to_frame())\n",
    "\n",
    "print(\"\\nLanguages\")\n",
    "display(df[\"language\"].value_counts().rename(\"count\").to_frame())\n",
    "\n",
    "print(\"\\nSecrets per sample\")\n",
    "display(df[\"sensitive_count\"].describe().to_frame(\"sensitive_count\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ax = df[\"category\"].value_counts().plot(kind=\"bar\", figsize=(8, 3), title=\"Samples by category\")\n",
    "ax.set_xlabel(\"category\")\n",
    "ax.set_ylabel(\"count\")\n",
    "ax.tick_params(axis=\"x\", rotation=45)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Preview one sample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "row = df.sample(1, random_state=42).iloc[0]\n",
    "print(\"id      :\", row[\"id\"])\n",
    "print(\"category:\", row[\"category\"])\n",
    "print(\"language:\", row[\"language\"])\n",
    "print(\"findings:\", row[\"finding_types\"])\n",
    "print(\"count   :\", row[\"sensitive_count\"])\n",
    "print(\"\\n--- code_text ---\\n\")\n",
    "# CSV may store newlines as the two characters \\\\n\n",
    "code = str(row[\"code_text\"]).replace(\"\\\\n\", \"\\n\")\n",
    "print(code)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Ground truth from JSON (recommended for evaluation)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "if json_path is not None and json_path.exists():\n",
    "    payload = json.loads(json_path.read_text(encoding=\"utf-8\"))\n",
    "    records = payload[\"records\"]\n",
    "    print(\"name :\", payload.get(\"name\"))\n",
    "    print(\"size :\", payload.get(\"size\"))\n",
    "    print(\"owasp:\", payload.get(\"owasp_primary\"))\n",
    "\n",
    "    sample = records[0]\n",
    "    print(\"\\nExample ground truth:\")\n",
    "    print(\"id      :\", sample[\"id\"])\n",
    "    print(\"findings:\")\n",
    "    for f in sample[\"sensitive_findings\"]:\n",
    "        print(\" -\", f[\"type\"], \"=\", f[\"value\"][:48] + (\"...\" if len(f[\"value\"]) > 48 else \"\"))\n",
    "else:\n",
    "    print(\"JSON not found in this runtime; use the CSV columns for a quick look.\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. Tiny detection baseline (substring match)\n",
    "\n",
    "Toy baseline: if a labeled secret string appears in `code_text`, count it as detected.  \n",
    "Replace this with your masking / detector pipeline."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "if json_path is not None and json_path.exists():\n",
    "    records = json.loads(json_path.read_text(encoding=\"utf-8\"))[\"records\"]\n",
    "    total = hit = 0\n",
    "    for r in records:\n",
    "        code = r[\"code_text\"]\n",
    "        for f in r[\"sensitive_findings\"]:\n",
    "            total += 1\n",
    "            if f[\"value\"] in code:\n",
    "                hit += 1\n",
    "    print(f\"Labeled secrets present in code_text: {hit}/{total} ({100 * hit / total:.1f}%)\")\n",
    "    print(\"(Expected ~100% — sanity check that labels match the snippets.)\")\n",
    "else:\n",
    "    print(\"Skip: JSON required for this cell.\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Next steps\n",
    "1. Build a detector (regex / ML) over `code_text`\n",
    "2. Compare predictions to `sensitive_findings`\n",
    "3. Measure precision, recall, and latency for your masking layer"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "name": "python",
   "version": "3.11.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}