(filePath, onRow, opts = {})
| 126 | } |
| 127 | |
| 128 | function forEachJsonLine(filePath, onRow, opts = {}) { |
| 129 | const stats = { parsed: 0, corrupt: 0, overlong: 0 }; |
| 130 | if (!fs.existsSync(filePath)) return stats; |
| 131 | |
| 132 | const maxLineBytes = Math.max(1, Math.floor(Number(opts.maxLineBytes) || resolveMaxJsonlLineBytes())); |
| 133 | const chunk = Buffer.allocUnsafe(JSONL_READ_CHUNK_BYTES); |
| 134 | let fd = null; |
| 135 | let parts = []; |
| 136 | let lineBytes = 0; |
| 137 | let dropping = false; |
| 138 | |
| 139 | const resetLine = () => { |
| 140 | parts = []; |
| 141 | lineBytes = 0; |
| 142 | dropping = false; |
| 143 | }; |
| 144 | |
| 145 | const appendSegment = (segment) => { |
| 146 | if (!segment || segment.length === 0 || dropping) return; |
| 147 | if (lineBytes + segment.length > maxLineBytes) { |
| 148 | parts = []; |
| 149 | lineBytes = 0; |
| 150 | dropping = true; |
| 151 | return; |
| 152 | } |
| 153 | parts.push(Buffer.from(segment)); |
| 154 | lineBytes += segment.length; |
| 155 | }; |
| 156 | |
| 157 | const finishLine = (hasNewline = false) => { |
| 158 | const totalLineBytes = lineBytes + (hasNewline ? 1 : 0); |
| 159 | if (dropping) { |
| 160 | stats.overlong += 1; |
| 161 | resetLine(); |
| 162 | return; |
| 163 | } |
| 164 | if (totalLineBytes > maxLineBytes) { |
| 165 | stats.overlong += 1; |
| 166 | resetLine(); |
| 167 | return; |
| 168 | } |
| 169 | if (lineBytes === 0) { |
| 170 | resetLine(); |
| 171 | return; |
| 172 | } |
| 173 | const trimmed = Buffer.concat(parts, lineBytes).toString('utf8').trim(); |
| 174 | resetLine(); |
| 175 | if (!trimmed) return; |
| 176 | try { |
| 177 | onRow(JSON.parse(trimmed)); |
| 178 | stats.parsed += 1; |
| 179 | } catch { |
| 180 | stats.corrupt += 1; |
| 181 | } |
| 182 | }; |
| 183 | |
| 184 | try { |
| 185 | fd = fs.openSync(filePath, 'r'); |
no test coverage detected