Spaces:
Sleeping
Sleeping
deploy at 2025-12-28 16:31:13.332600
Browse files- Dependency length.ipynb +160 -364
- main.py +10 -0
Dependency length.ipynb
CHANGED
|
@@ -38,7 +38,7 @@
|
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"cell_type": "code",
|
| 41 |
-
"execution_count":
|
| 42 |
"id": "cefdbf22-b747-4bea-b279-c9b01e75ff2e",
|
| 43 |
"metadata": {},
|
| 44 |
"outputs": [],
|
|
@@ -62,6 +62,11 @@
|
|
| 62 |
" to_reverse[i-1].dep_ = \"comp\"\n",
|
| 63 |
" for token in doc:\n",
|
| 64 |
" if token.dep_ == \"dep\": token.dep_ = \"unknown\"\n",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 65 |
" for rel in relations.keys():\n",
|
| 66 |
" if token.dep_ in relations[rel]: token.dep_ = rel\n",
|
| 67 |
" for token in doc:\n",
|
|
@@ -75,12 +80,16 @@
|
|
| 75 |
" for child in [c for c in token.children if c.i < subject[0].i]:\n",
|
| 76 |
" if child.dep_ in [\"comp\", \"udep\"]:\n",
|
| 77 |
" child.dep_ = \"mod\"\n",
|
|
|
|
|
|
|
|
|
|
|
|
|
| 78 |
" return doc"
|
| 79 |
]
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"cell_type": "code",
|
| 83 |
-
"execution_count":
|
| 84 |
"id": "23efda66-9d58-4169-9fa0-05de47267b5a",
|
| 85 |
"metadata": {},
|
| 86 |
"outputs": [],
|
|
@@ -98,7 +107,7 @@
|
|
| 98 |
},
|
| 99 |
{
|
| 100 |
"cell_type": "code",
|
| 101 |
-
"execution_count":
|
| 102 |
"id": "5bb29f0a-9d55-4f2d-84c3-4c3c4bae0271",
|
| 103 |
"metadata": {},
|
| 104 |
"outputs": [],
|
|
@@ -132,472 +141,264 @@
|
|
| 132 |
},
|
| 133 |
{
|
| 134 |
"cell_type": "code",
|
| 135 |
-
"execution_count":
|
| 136 |
"id": "ba90ff19-c665-49d8-8ad4-5caee885901d",
|
| 137 |
"metadata": {},
|
| 138 |
"outputs": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 139 |
{
|
| 140 |
"data": {
|
| 141 |
"text/html": [
|
| 142 |
-
"<span class=\"tex2jax_ignore\"><svg xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\" xml:lang=\"en\" id=\"
|
| 143 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 144 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"50\">
|
| 145 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"50\">
|
| 146 |
-
"</text>\n",
|
| 147 |
-
"\n",
|
| 148 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 149 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"225\">to</tspan>\n",
|
| 150 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"225\">ADP</tspan>\n",
|
| 151 |
-
"</text>\n",
|
| 152 |
-
"\n",
|
| 153 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 154 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"400\">the</tspan>\n",
|
| 155 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"400\">DET</tspan>\n",
|
| 156 |
-
"</text>\n",
|
| 157 |
-
"\n",
|
| 158 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 159 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"575\">studies, “</tspan>\n",
|
| 160 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"575\">NOUN</tspan>\n",
|
| 161 |
-
"</text>\n",
|
| 162 |
-
"\n",
|
| 163 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 164 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"750\">People</tspan>\n",
|
| 165 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"750\">NOUN</tspan>\n",
|
| 166 |
-
"</text>\n",
|
| 167 |
-
"\n",
|
| 168 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 169 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"925\">who</tspan>\n",
|
| 170 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"925\">PRON</tspan>\n",
|
| 171 |
-
"</text>\n",
|
| 172 |
-
"\n",
|
| 173 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 174 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"1100\">danced</tspan>\n",
|
| 175 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"1100\">VERB</tspan>\n",
|
| 176 |
-
"</text>\n",
|
| 177 |
-
"\n",
|
| 178 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 179 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"1275\">more</tspan>\n",
|
| 180 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"1275\">ADJ</tspan>\n",
|
| 181 |
-
"</text>\n",
|
| 182 |
-
"\n",
|
| 183 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 184 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"1450\">than</tspan>\n",
|
| 185 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"1450\">ADP</tspan>\n",
|
| 186 |
-
"</text>\n",
|
| 187 |
-
"\n",
|
| 188 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 189 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"1625\">once</tspan>\n",
|
| 190 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"1625\">ADV</tspan>\n",
|
| 191 |
-
"</text>\n",
|
| 192 |
-
"\n",
|
| 193 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 194 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"1800\">a</tspan>\n",
|
| 195 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"1800\">DET</tspan>\n",
|
| 196 |
-
"</text>\n",
|
| 197 |
-
"\n",
|
| 198 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 199 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"1975\">week</tspan>\n",
|
| 200 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"1975\">NOUN</tspan>\n",
|
| 201 |
-
"</text>\n",
|
| 202 |
-
"\n",
|
| 203 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 204 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"2150\">had</tspan>\n",
|
| 205 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"2150\">VERB</tspan>\n",
|
| 206 |
-
"</text>\n",
|
| 207 |
-
"\n",
|
| 208 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 209 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"2325\">a</tspan>\n",
|
| 210 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"2325\">DET</tspan>\n",
|
| 211 |
-
"</text>\n",
|
| 212 |
-
"\n",
|
| 213 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 214 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"2500\">76%</tspan>\n",
|
| 215 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"2500\">NOUN</tspan>\n",
|
| 216 |
-
"</text>\n",
|
| 217 |
-
"\n",
|
| 218 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 219 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"2675\">lower</tspan>\n",
|
| 220 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"2675\">ADJ</tspan>\n",
|
| 221 |
-
"</text>\n",
|
| 222 |
-
"\n",
|
| 223 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 224 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"2850\">risk</tspan>\n",
|
| 225 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"2850\">NOUN</tspan>\n",
|
| 226 |
"</text>\n",
|
| 227 |
"\n",
|
| 228 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 229 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 230 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 231 |
"</text>\n",
|
| 232 |
"\n",
|
| 233 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 234 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 235 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 236 |
"</text>\n",
|
| 237 |
"\n",
|
| 238 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 239 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 240 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 241 |
"</text>\n",
|
| 242 |
"\n",
|
| 243 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 244 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 245 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 246 |
"</text>\n",
|
| 247 |
"\n",
|
| 248 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 249 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 250 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 251 |
"</text>\n",
|
| 252 |
"\n",
|
| 253 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 254 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 255 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 256 |
"</text>\n",
|
| 257 |
"\n",
|
| 258 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 259 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 260 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 261 |
"</text>\n",
|
| 262 |
"\n",
|
| 263 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 264 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 265 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 266 |
"</text>\n",
|
| 267 |
"\n",
|
| 268 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 269 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 270 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 271 |
"</text>\n",
|
| 272 |
"\n",
|
| 273 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 274 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 275 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 276 |
"</text>\n",
|
| 277 |
"\n",
|
| 278 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 279 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 280 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 281 |
"</text>\n",
|
| 282 |
"\n",
|
| 283 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 284 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 285 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 286 |
"</text>\n",
|
| 287 |
"\n",
|
| 288 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 289 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 290 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 291 |
"</text>\n",
|
| 292 |
"\n",
|
| 293 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 294 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 295 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 296 |
"</text>\n",
|
| 297 |
"\n",
|
| 298 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 299 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 300 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 301 |
"</text>\n",
|
| 302 |
"\n",
|
| 303 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 304 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 305 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 306 |
"</text>\n",
|
| 307 |
"\n",
|
| 308 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"
|
| 309 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"
|
| 310 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"
|
| 311 |
"</text>\n",
|
| 312 |
"\n",
|
| 313 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 314 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"6000\">Pasricha</tspan>\n",
|
| 315 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"6000\">PROPN</tspan>\n",
|
| 316 |
-
"</text>\n",
|
| 317 |
-
"\n",
|
| 318 |
-
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"484.5\">\n",
|
| 319 |
-
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"6175\">noted.</tspan>\n",
|
| 320 |
-
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"6175\">VERB</tspan>\n",
|
| 321 |
-
"</text>\n",
|
| 322 |
-
"\n",
|
| 323 |
-
"<g class=\"displacy-arrow\">\n",
|
| 324 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-0\" stroke-width=\"2px\" d=\"M70,439.5 C70,89.5 2145.0,89.5 2145.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 325 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 326 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-0\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">mod</textPath>\n",
|
| 327 |
-
" </text>\n",
|
| 328 |
-
" <path class=\"displacy-arrowhead\" d=\"M70,441.5 L62,429.5 78,429.5\" fill=\"currentColor\"/>\n",
|
| 329 |
-
"</g>\n",
|
| 330 |
-
"\n",
|
| 331 |
-
"<g class=\"displacy-arrow\">\n",
|
| 332 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-1\" stroke-width=\"2px\" d=\"M70,439.5 C70,352.0 205.0,352.0 205.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 333 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 334 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-1\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">udep</textPath>\n",
|
| 335 |
-
" </text>\n",
|
| 336 |
-
" <path class=\"displacy-arrowhead\" d=\"M205.0,441.5 L213.0,429.5 197.0,429.5\" fill=\"currentColor\"/>\n",
|
| 337 |
-
"</g>\n",
|
| 338 |
-
"\n",
|
| 339 |
-
"<g class=\"displacy-arrow\">\n",
|
| 340 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-2\" stroke-width=\"2px\" d=\"M420,439.5 C420,352.0 555.0,352.0 555.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 341 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 342 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-2\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">det</textPath>\n",
|
| 343 |
-
" </text>\n",
|
| 344 |
-
" <path class=\"displacy-arrowhead\" d=\"M420,441.5 L412,429.5 428,429.5\" fill=\"currentColor\"/>\n",
|
| 345 |
-
"</g>\n",
|
| 346 |
-
"\n",
|
| 347 |
-
"<g class=\"displacy-arrow\">\n",
|
| 348 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-3\" stroke-width=\"2px\" d=\"M245,439.5 C245,264.5 560.0,264.5 560.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 349 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 350 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-3\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">comp</textPath>\n",
|
| 351 |
-
" </text>\n",
|
| 352 |
-
" <path class=\"displacy-arrowhead\" d=\"M560.0,441.5 L568.0,429.5 552.0,429.5\" fill=\"currentColor\"/>\n",
|
| 353 |
-
"</g>\n",
|
| 354 |
-
"\n",
|
| 355 |
-
"<g class=\"displacy-arrow\">\n",
|
| 356 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-4\" stroke-width=\"2px\" d=\"M770,439.5 C770,177.0 2140.0,177.0 2140.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 357 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 358 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-4\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">subj</textPath>\n",
|
| 359 |
-
" </text>\n",
|
| 360 |
-
" <path class=\"displacy-arrowhead\" d=\"M770,441.5 L762,429.5 778,429.5\" fill=\"currentColor\"/>\n",
|
| 361 |
-
"</g>\n",
|
| 362 |
-
"\n",
|
| 363 |
-
"<g class=\"displacy-arrow\">\n",
|
| 364 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-5\" stroke-width=\"2px\" d=\"M945,439.5 C945,352.0 1080.0,352.0 1080.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 365 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 366 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-5\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">subj</textPath>\n",
|
| 367 |
-
" </text>\n",
|
| 368 |
-
" <path class=\"displacy-arrowhead\" d=\"M945,441.5 L937,429.5 953,429.5\" fill=\"currentColor\"/>\n",
|
| 369 |
-
"</g>\n",
|
| 370 |
-
"\n",
|
| 371 |
-
"<g class=\"displacy-arrow\">\n",
|
| 372 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-6\" stroke-width=\"2px\" d=\"M770,439.5 C770,264.5 1085.0,264.5 1085.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 373 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 374 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-6\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">mod</textPath>\n",
|
| 375 |
-
" </text>\n",
|
| 376 |
-
" <path class=\"displacy-arrowhead\" d=\"M1085.0,441.5 L1093.0,429.5 1077.0,429.5\" fill=\"currentColor\"/>\n",
|
| 377 |
-
"</g>\n",
|
| 378 |
-
"\n",
|
| 379 |
-
"<g class=\"displacy-arrow\">\n",
|
| 380 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-7\" stroke-width=\"2px\" d=\"M1120,439.5 C1120,352.0 1255.0,352.0 1255.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 381 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 382 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-7\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">comp</textPath>\n",
|
| 383 |
-
" </text>\n",
|
| 384 |
-
" <path class=\"displacy-arrowhead\" d=\"M1255.0,441.5 L1263.0,429.5 1247.0,429.5\" fill=\"currentColor\"/>\n",
|
| 385 |
-
"</g>\n",
|
| 386 |
-
"\n",
|
| 387 |
-
"<g class=\"displacy-arrow\">\n",
|
| 388 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-8\" stroke-width=\"2px\" d=\"M1295,439.5 C1295,352.0 1430.0,352.0 1430.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 389 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 390 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-8\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">udep</textPath>\n",
|
| 391 |
-
" </text>\n",
|
| 392 |
-
" <path class=\"displacy-arrowhead\" d=\"M1430.0,441.5 L1438.0,429.5 1422.0,429.5\" fill=\"currentColor\"/>\n",
|
| 393 |
-
"</g>\n",
|
| 394 |
-
"\n",
|
| 395 |
-
"<g class=\"displacy-arrow\">\n",
|
| 396 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-9\" stroke-width=\"2px\" d=\"M1470,439.5 C1470,352.0 1605.0,352.0 1605.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 397 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 398 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-9\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">pcomp</textPath>\n",
|
| 399 |
-
" </text>\n",
|
| 400 |
-
" <path class=\"displacy-arrowhead\" d=\"M1605.0,441.5 L1613.0,429.5 1597.0,429.5\" fill=\"currentColor\"/>\n",
|
| 401 |
-
"</g>\n",
|
| 402 |
-
"\n",
|
| 403 |
-
"<g class=\"displacy-arrow\">\n",
|
| 404 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-10\" stroke-width=\"2px\" d=\"M1820,439.5 C1820,352.0 1955.0,352.0 1955.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 405 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 406 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-10\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">det</textPath>\n",
|
| 407 |
-
" </text>\n",
|
| 408 |
-
" <path class=\"displacy-arrowhead\" d=\"M1820,441.5 L1812,429.5 1828,429.5\" fill=\"currentColor\"/>\n",
|
| 409 |
-
"</g>\n",
|
| 410 |
-
"\n",
|
| 411 |
-
"<g class=\"displacy-arrow\">\n",
|
| 412 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-11\" stroke-width=\"2px\" d=\"M1295,439.5 C1295,264.5 1960.0,264.5 1960.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 413 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 414 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-11\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">mod</textPath>\n",
|
| 415 |
-
" </text>\n",
|
| 416 |
-
" <path class=\"displacy-arrowhead\" d=\"M1960.0,441.5 L1968.0,429.5 1952.0,429.5\" fill=\"currentColor\"/>\n",
|
| 417 |
-
"</g>\n",
|
| 418 |
-
"\n",
|
| 419 |
-
"<g class=\"displacy-arrow\">\n",
|
| 420 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-12\" stroke-width=\"2px\" d=\"M2170,439.5 C2170,2.0 6175.0,2.0 6175.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 421 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 422 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-12\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">mod</textPath>\n",
|
| 423 |
-
" </text>\n",
|
| 424 |
-
" <path class=\"displacy-arrowhead\" d=\"M2170,441.5 L2162,429.5 2178,429.5\" fill=\"currentColor\"/>\n",
|
| 425 |
-
"</g>\n",
|
| 426 |
-
"\n",
|
| 427 |
-
"<g class=\"displacy-arrow\">\n",
|
| 428 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-13\" stroke-width=\"2px\" d=\"M2345,439.5 C2345,264.5 2835.0,264.5 2835.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 429 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 430 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-13\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">det</textPath>\n",
|
| 431 |
-
" </text>\n",
|
| 432 |
-
" <path class=\"displacy-arrowhead\" d=\"M2345,441.5 L2337,429.5 2353,429.5\" fill=\"currentColor\"/>\n",
|
| 433 |
-
"</g>\n",
|
| 434 |
-
"\n",
|
| 435 |
-
"<g class=\"displacy-arrow\">\n",
|
| 436 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-14\" stroke-width=\"2px\" d=\"M2520,439.5 C2520,352.0 2655.0,352.0 2655.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 437 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 438 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-14\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">mod</textPath>\n",
|
| 439 |
-
" </text>\n",
|
| 440 |
-
" <path class=\"displacy-arrowhead\" d=\"M2520,441.5 L2512,429.5 2528,429.5\" fill=\"currentColor\"/>\n",
|
| 441 |
-
"</g>\n",
|
| 442 |
-
"\n",
|
| 443 |
"<g class=\"displacy-arrow\">\n",
|
| 444 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 445 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 446 |
-
" <textPath xlink:href=\"#arrow-
|
| 447 |
" </text>\n",
|
| 448 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 449 |
"</g>\n",
|
| 450 |
"\n",
|
| 451 |
"<g class=\"displacy-arrow\">\n",
|
| 452 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 453 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 454 |
-
" <textPath xlink:href=\"#arrow-
|
| 455 |
" </text>\n",
|
| 456 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 457 |
"</g>\n",
|
| 458 |
"\n",
|
| 459 |
"<g class=\"displacy-arrow\">\n",
|
| 460 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 461 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 462 |
-
" <textPath xlink:href=\"#arrow-
|
| 463 |
" </text>\n",
|
| 464 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 465 |
"</g>\n",
|
| 466 |
"\n",
|
| 467 |
"<g class=\"displacy-arrow\">\n",
|
| 468 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 469 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 470 |
-
" <textPath xlink:href=\"#arrow-
|
| 471 |
" </text>\n",
|
| 472 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 473 |
"</g>\n",
|
| 474 |
"\n",
|
| 475 |
"<g class=\"displacy-arrow\">\n",
|
| 476 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 477 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 478 |
-
" <textPath xlink:href=\"#arrow-
|
| 479 |
" </text>\n",
|
| 480 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 481 |
"</g>\n",
|
| 482 |
"\n",
|
| 483 |
"<g class=\"displacy-arrow\">\n",
|
| 484 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 485 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 486 |
-
" <textPath xlink:href=\"#arrow-
|
| 487 |
" </text>\n",
|
| 488 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 489 |
"</g>\n",
|
| 490 |
"\n",
|
| 491 |
"<g class=\"displacy-arrow\">\n",
|
| 492 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 493 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 494 |
-
" <textPath xlink:href=\"#arrow-
|
| 495 |
" </text>\n",
|
| 496 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 497 |
"</g>\n",
|
| 498 |
"\n",
|
| 499 |
"<g class=\"displacy-arrow\">\n",
|
| 500 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 501 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 502 |
-
" <textPath xlink:href=\"#arrow-
|
| 503 |
" </text>\n",
|
| 504 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 505 |
"</g>\n",
|
| 506 |
"\n",
|
| 507 |
"<g class=\"displacy-arrow\">\n",
|
| 508 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 509 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 510 |
-
" <textPath xlink:href=\"#arrow-
|
| 511 |
" </text>\n",
|
| 512 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 513 |
"</g>\n",
|
| 514 |
"\n",
|
| 515 |
"<g class=\"displacy-arrow\">\n",
|
| 516 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 517 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 518 |
-
" <textPath xlink:href=\"#arrow-
|
| 519 |
" </text>\n",
|
| 520 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 521 |
"</g>\n",
|
| 522 |
"\n",
|
| 523 |
"<g class=\"displacy-arrow\">\n",
|
| 524 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 525 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 526 |
-
" <textPath xlink:href=\"#arrow-
|
| 527 |
" </text>\n",
|
| 528 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 529 |
"</g>\n",
|
| 530 |
"\n",
|
| 531 |
"<g class=\"displacy-arrow\">\n",
|
| 532 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 533 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 534 |
-
" <textPath xlink:href=\"#arrow-
|
| 535 |
" </text>\n",
|
| 536 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 537 |
"</g>\n",
|
| 538 |
"\n",
|
| 539 |
"<g class=\"displacy-arrow\">\n",
|
| 540 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 541 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 542 |
-
" <textPath xlink:href=\"#arrow-
|
| 543 |
" </text>\n",
|
| 544 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 545 |
"</g>\n",
|
| 546 |
"\n",
|
| 547 |
"<g class=\"displacy-arrow\">\n",
|
| 548 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 549 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 550 |
-
" <textPath xlink:href=\"#arrow-
|
| 551 |
" </text>\n",
|
| 552 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 553 |
"</g>\n",
|
| 554 |
"\n",
|
| 555 |
"<g class=\"displacy-arrow\">\n",
|
| 556 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 557 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 558 |
-
" <textPath xlink:href=\"#arrow-
|
| 559 |
" </text>\n",
|
| 560 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 561 |
"</g>\n",
|
| 562 |
"\n",
|
| 563 |
"<g class=\"displacy-arrow\">\n",
|
| 564 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 565 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 566 |
-
" <textPath xlink:href=\"#arrow-
|
| 567 |
" </text>\n",
|
| 568 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 569 |
"</g>\n",
|
| 570 |
"\n",
|
| 571 |
"<g class=\"displacy-arrow\">\n",
|
| 572 |
-
" <path class=\"displacy-arc\" id=\"arrow-
|
| 573 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 574 |
-
" <textPath xlink:href=\"#arrow-
|
| 575 |
" </text>\n",
|
| 576 |
-
" <path class=\"displacy-arrowhead\" d=\"
|
| 577 |
-
"</g>\n",
|
| 578 |
-
"\n",
|
| 579 |
-
"<g class=\"displacy-arrow\">\n",
|
| 580 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-32\" stroke-width=\"2px\" d=\"M4970,439.5 C4970,177.0 5640.0,177.0 5640.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 581 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 582 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-32\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">mod</textPath>\n",
|
| 583 |
-
" </text>\n",
|
| 584 |
-
" <path class=\"displacy-arrowhead\" d=\"M5640.0,441.5 L5648.0,429.5 5632.0,429.5\" fill=\"currentColor\"/>\n",
|
| 585 |
-
"</g>\n",
|
| 586 |
-
"\n",
|
| 587 |
-
"<g class=\"displacy-arrow\">\n",
|
| 588 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-33\" stroke-width=\"2px\" d=\"M5845,439.5 C5845,352.0 5980.0,352.0 5980.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 589 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 590 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-33\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">compound</textPath>\n",
|
| 591 |
-
" </text>\n",
|
| 592 |
-
" <path class=\"displacy-arrowhead\" d=\"M5845,441.5 L5837,429.5 5853,429.5\" fill=\"currentColor\"/>\n",
|
| 593 |
-
"</g>\n",
|
| 594 |
-
"\n",
|
| 595 |
-
"<g class=\"displacy-arrow\">\n",
|
| 596 |
-
" <path class=\"displacy-arc\" id=\"arrow-0c70d09209f64114aa5a12db7f67e2b2-0-34\" stroke-width=\"2px\" d=\"M6020,439.5 C6020,352.0 6155.0,352.0 6155.0,439.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 597 |
-
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 598 |
-
" <textPath xlink:href=\"#arrow-0c70d09209f64114aa5a12db7f67e2b2-0-34\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">subj</textPath>\n",
|
| 599 |
-
" </text>\n",
|
| 600 |
-
" <path class=\"displacy-arrowhead\" d=\"M6020,441.5 L6012,429.5 6028,429.5\" fill=\"currentColor\"/>\n",
|
| 601 |
"</g>\n",
|
| 602 |
"</svg></span>"
|
| 603 |
],
|
|
@@ -610,24 +411,30 @@
|
|
| 610 |
}
|
| 611 |
],
|
| 612 |
"source": [
|
| 613 |
-
"doc = ssudify(nlp(\"
|
| 614 |
"# Since this is an interactive Jupyter environment, we can use displacy.render here\n",
|
| 615 |
"displacy.render(doc, style='dep')"
|
| 616 |
]
|
| 617 |
},
|
| 618 |
{
|
| 619 |
"cell_type": "code",
|
| 620 |
-
"execution_count":
|
| 621 |
"id": "daa1e8f5-2afe-4153-9469-a4c2af9b46b1",
|
| 622 |
"metadata": {},
|
| 623 |
"outputs": [
|
| 624 |
{
|
| 625 |
"data": {
|
| 626 |
"text/plain": [
|
| 627 |
-
"[(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 628 |
]
|
| 629 |
},
|
| 630 |
-
"execution_count":
|
| 631 |
"metadata": {},
|
| 632 |
"output_type": "execute_result"
|
| 633 |
}
|
|
@@ -650,17 +457,6 @@
|
|
| 650 |
{
|
| 651 |
"cell_type": "code",
|
| 652 |
"execution_count": 9,
|
| 653 |
-
"id": "9f23a19c-b645-4b78-85fd-bc9742d00f82",
|
| 654 |
-
"metadata": {},
|
| 655 |
-
"outputs": [],
|
| 656 |
-
"source": [
|
| 657 |
-
"def display_annotations(annot):\n",
|
| 658 |
-
" return Div(*[Span(a[0], style=f\"margin-right: 0.5ex; background: rgba(219, 144, 232, {a[1]/15})\") for a in annot])"
|
| 659 |
-
]
|
| 660 |
-
},
|
| 661 |
-
{
|
| 662 |
-
"cell_type": "code",
|
| 663 |
-
"execution_count": 10,
|
| 664 |
"id": "8fb35fe3-68de-45b6-8053-8b390c8e8661",
|
| 665 |
"metadata": {},
|
| 666 |
"outputs": [],
|
|
@@ -670,7 +466,7 @@
|
|
| 670 |
},
|
| 671 |
{
|
| 672 |
"cell_type": "code",
|
| 673 |
-
"execution_count":
|
| 674 |
"id": "9ce9c318-733d-4ee1-b5d7-39d98aa14697",
|
| 675 |
"metadata": {},
|
| 676 |
"outputs": [],
|
|
@@ -708,7 +504,7 @@
|
|
| 708 |
},
|
| 709 |
{
|
| 710 |
"cell_type": "code",
|
| 711 |
-
"execution_count":
|
| 712 |
"id": "4cc68772-d26d-4cd1-be59-c23a5739c3d0",
|
| 713 |
"metadata": {},
|
| 714 |
"outputs": [
|
|
@@ -738,7 +534,7 @@
|
|
| 738 |
},
|
| 739 |
{
|
| 740 |
"cell_type": "code",
|
| 741 |
-
"execution_count":
|
| 742 |
"id": "3e8c0a41-0d79-4715-ab2b-e5a2d0455569",
|
| 743 |
"metadata": {},
|
| 744 |
"outputs": [
|
|
@@ -757,7 +553,7 @@
|
|
| 757 |
"<IPython.core.display.HTML object>"
|
| 758 |
]
|
| 759 |
},
|
| 760 |
-
"execution_count":
|
| 761 |
"metadata": {},
|
| 762 |
"output_type": "execute_result"
|
| 763 |
}
|
|
@@ -793,7 +589,7 @@
|
|
| 793 |
"name": "python",
|
| 794 |
"nbconvert_exporter": "python",
|
| 795 |
"pygments_lexer": "ipython3",
|
| 796 |
-
"version": "3.
|
| 797 |
}
|
| 798 |
},
|
| 799 |
"nbformat": 4,
|
|
|
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"cell_type": "code",
|
| 41 |
+
"execution_count": 56,
|
| 42 |
"id": "cefdbf22-b747-4bea-b279-c9b01e75ff2e",
|
| 43 |
"metadata": {},
|
| 44 |
"outputs": [],
|
|
|
|
| 62 |
" to_reverse[i-1].dep_ = \"comp\"\n",
|
| 63 |
" for token in doc:\n",
|
| 64 |
" if token.dep_ == \"dep\": token.dep_ = \"unknown\"\n",
|
| 65 |
+
" if token.dep_ == \"prep\" and token.head.pos_ == \"VERB\" and token.i < token.head.i:\n",
|
| 66 |
+
" token.dep_ = \"mod\"\n",
|
| 67 |
+
" if token.dep_ == \"ccomp\" and any(sibling.dep_ in relations[\"comp\"] \n",
|
| 68 |
+
" for sibling in token.head.rights if sibling.i < token.i):\n",
|
| 69 |
+
" token.dep_ = \"mod\"\n",
|
| 70 |
" for rel in relations.keys():\n",
|
| 71 |
" if token.dep_ in relations[rel]: token.dep_ = rel\n",
|
| 72 |
" for token in doc:\n",
|
|
|
|
| 80 |
" for child in [c for c in token.children if c.i < subject[0].i]:\n",
|
| 81 |
" if child.dep_ in [\"comp\", \"udep\"]:\n",
|
| 82 |
" child.dep_ = \"mod\"\n",
|
| 83 |
+
" for token in doc:\n",
|
| 84 |
+
" if any(t.text == \";\" for t in doc if (token.i < t.i < token.head.i or token.head.i < t.i < token.i) and token.pos_ != \"PUNCT\"):\n",
|
| 85 |
+
" token.head = token\n",
|
| 86 |
+
" token.dep_ = \"root\"\n",
|
| 87 |
" return doc"
|
| 88 |
]
|
| 89 |
},
|
| 90 |
{
|
| 91 |
"cell_type": "code",
|
| 92 |
+
"execution_count": 28,
|
| 93 |
"id": "23efda66-9d58-4169-9fa0-05de47267b5a",
|
| 94 |
"metadata": {},
|
| 95 |
"outputs": [],
|
|
|
|
| 107 |
},
|
| 108 |
{
|
| 109 |
"cell_type": "code",
|
| 110 |
+
"execution_count": 5,
|
| 111 |
"id": "5bb29f0a-9d55-4f2d-84c3-4c3c4bae0271",
|
| 112 |
"metadata": {},
|
| 113 |
"outputs": [],
|
|
|
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"cell_type": "code",
|
| 144 |
+
"execution_count": 57,
|
| 145 |
"id": "ba90ff19-c665-49d8-8ad4-5caee885901d",
|
| 146 |
"metadata": {},
|
| 147 |
"outputs": [
|
| 148 |
+
{
|
| 149 |
+
"name": "stdout",
|
| 150 |
+
"output_type": "stream",
|
| 151 |
+
"text": [
|
| 152 |
+
"In []\n",
|
| 153 |
+
"1849 []\n",
|
| 154 |
+
"he []\n",
|
| 155 |
+
"began []\n",
|
| 156 |
+
"public []\n",
|
| 157 |
+
"readings []\n",
|
| 158 |
+
"of []\n",
|
| 159 |
+
"the []\n",
|
| 160 |
+
"story []\n",
|
| 161 |
+
", []\n",
|
| 162 |
+
"which []\n",
|
| 163 |
+
"proved ['punct']\n",
|
| 164 |
+
"so []\n",
|
| 165 |
+
"successful []\n",
|
| 166 |
+
"he []\n",
|
| 167 |
+
"undertook ['comp']\n",
|
| 168 |
+
"127 []\n",
|
| 169 |
+
"further []\n",
|
| 170 |
+
"performances []\n",
|
| 171 |
+
". ['comp']\n"
|
| 172 |
+
]
|
| 173 |
+
},
|
| 174 |
{
|
| 175 |
"data": {
|
| 176 |
"text/html": [
|
| 177 |
+
"<span class=\"tex2jax_ignore\"><svg xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\" xml:lang=\"en\" id=\"5232fcca8f23464a9783eb067e24fe32-0\" class=\"displacy\" width=\"3200\" height=\"399.5\" direction=\"ltr\" style=\"max-width: none; height: 399.5px; color: #000000; background: #ffffff; font-family: Arial; direction: ltr\">\n",
|
| 178 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 179 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"50\">In</tspan>\n",
|
| 180 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"50\">ADP</tspan>\n",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 181 |
"</text>\n",
|
| 182 |
"\n",
|
| 183 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 184 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"225\">1849</tspan>\n",
|
| 185 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"225\">NUM</tspan>\n",
|
| 186 |
"</text>\n",
|
| 187 |
"\n",
|
| 188 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 189 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"400\">he</tspan>\n",
|
| 190 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"400\">PRON</tspan>\n",
|
| 191 |
"</text>\n",
|
| 192 |
"\n",
|
| 193 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 194 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"575\">began</tspan>\n",
|
| 195 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"575\">VERB</tspan>\n",
|
| 196 |
"</text>\n",
|
| 197 |
"\n",
|
| 198 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 199 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"750\">public</tspan>\n",
|
| 200 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"750\">ADJ</tspan>\n",
|
| 201 |
"</text>\n",
|
| 202 |
"\n",
|
| 203 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 204 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"925\">readings</tspan>\n",
|
| 205 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"925\">NOUN</tspan>\n",
|
| 206 |
"</text>\n",
|
| 207 |
"\n",
|
| 208 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 209 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"1100\">of</tspan>\n",
|
| 210 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"1100\">ADP</tspan>\n",
|
| 211 |
"</text>\n",
|
| 212 |
"\n",
|
| 213 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 214 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"1275\">the</tspan>\n",
|
| 215 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"1275\">DET</tspan>\n",
|
| 216 |
"</text>\n",
|
| 217 |
"\n",
|
| 218 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 219 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"1450\">story,</tspan>\n",
|
| 220 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"1450\">NOUN</tspan>\n",
|
| 221 |
"</text>\n",
|
| 222 |
"\n",
|
| 223 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 224 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"1625\">which</tspan>\n",
|
| 225 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"1625\">PRON</tspan>\n",
|
| 226 |
"</text>\n",
|
| 227 |
"\n",
|
| 228 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 229 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"1800\">proved</tspan>\n",
|
| 230 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"1800\">VERB</tspan>\n",
|
| 231 |
"</text>\n",
|
| 232 |
"\n",
|
| 233 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 234 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"1975\">so</tspan>\n",
|
| 235 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"1975\">ADV</tspan>\n",
|
| 236 |
"</text>\n",
|
| 237 |
"\n",
|
| 238 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 239 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"2150\">successful</tspan>\n",
|
| 240 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"2150\">ADJ</tspan>\n",
|
| 241 |
"</text>\n",
|
| 242 |
"\n",
|
| 243 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 244 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"2325\">he</tspan>\n",
|
| 245 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"2325\">PRON</tspan>\n",
|
| 246 |
"</text>\n",
|
| 247 |
"\n",
|
| 248 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 249 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"2500\">undertook</tspan>\n",
|
| 250 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"2500\">VERB</tspan>\n",
|
| 251 |
"</text>\n",
|
| 252 |
"\n",
|
| 253 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 254 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"2675\">127</tspan>\n",
|
| 255 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"2675\">NUM</tspan>\n",
|
| 256 |
"</text>\n",
|
| 257 |
"\n",
|
| 258 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 259 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"2850\">further</tspan>\n",
|
| 260 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"2850\">ADJ</tspan>\n",
|
| 261 |
"</text>\n",
|
| 262 |
"\n",
|
| 263 |
+
"<text class=\"displacy-token\" fill=\"currentColor\" text-anchor=\"middle\" y=\"309.5\">\n",
|
| 264 |
+
" <tspan class=\"displacy-word\" fill=\"currentColor\" x=\"3025\">performances.</tspan>\n",
|
| 265 |
+
" <tspan class=\"displacy-tag\" dy=\"2em\" fill=\"currentColor\" x=\"3025\">NOUN</tspan>\n",
|
| 266 |
"</text>\n",
|
| 267 |
"\n",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 268 |
"<g class=\"displacy-arrow\">\n",
|
| 269 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-0\" stroke-width=\"2px\" d=\"M70,264.5 C70,89.5 570.0,89.5 570.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 270 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 271 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-0\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">mod</textPath>\n",
|
| 272 |
" </text>\n",
|
| 273 |
+
" <path class=\"displacy-arrowhead\" d=\"M70,266.5 L62,254.5 78,254.5\" fill=\"currentColor\"/>\n",
|
| 274 |
"</g>\n",
|
| 275 |
"\n",
|
| 276 |
"<g class=\"displacy-arrow\">\n",
|
| 277 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-1\" stroke-width=\"2px\" d=\"M70,264.5 C70,177.0 215.0,177.0 215.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 278 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 279 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-1\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">comp</textPath>\n",
|
| 280 |
" </text>\n",
|
| 281 |
+
" <path class=\"displacy-arrowhead\" d=\"M215.0,266.5 L223.0,254.5 207.0,254.5\" fill=\"currentColor\"/>\n",
|
| 282 |
"</g>\n",
|
| 283 |
"\n",
|
| 284 |
"<g class=\"displacy-arrow\">\n",
|
| 285 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-2\" stroke-width=\"2px\" d=\"M420,264.5 C420,177.0 565.0,177.0 565.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 286 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 287 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-2\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">subj</textPath>\n",
|
| 288 |
" </text>\n",
|
| 289 |
+
" <path class=\"displacy-arrowhead\" d=\"M420,266.5 L412,254.5 428,254.5\" fill=\"currentColor\"/>\n",
|
| 290 |
"</g>\n",
|
| 291 |
"\n",
|
| 292 |
"<g class=\"displacy-arrow\">\n",
|
| 293 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-3\" stroke-width=\"2px\" d=\"M770,264.5 C770,177.0 915.0,177.0 915.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 294 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 295 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-3\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">udep</textPath>\n",
|
| 296 |
" </text>\n",
|
| 297 |
+
" <path class=\"displacy-arrowhead\" d=\"M770,266.5 L762,254.5 778,254.5\" fill=\"currentColor\"/>\n",
|
| 298 |
"</g>\n",
|
| 299 |
"\n",
|
| 300 |
"<g class=\"displacy-arrow\">\n",
|
| 301 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-4\" stroke-width=\"2px\" d=\"M595,264.5 C595,89.5 920.0,89.5 920.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 302 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 303 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-4\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">comp</textPath>\n",
|
| 304 |
" </text>\n",
|
| 305 |
+
" <path class=\"displacy-arrowhead\" d=\"M920.0,266.5 L928.0,254.5 912.0,254.5\" fill=\"currentColor\"/>\n",
|
| 306 |
"</g>\n",
|
| 307 |
"\n",
|
| 308 |
"<g class=\"displacy-arrow\">\n",
|
| 309 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-5\" stroke-width=\"2px\" d=\"M945,264.5 C945,177.0 1090.0,177.0 1090.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 310 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 311 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-5\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">udep</textPath>\n",
|
| 312 |
" </text>\n",
|
| 313 |
+
" <path class=\"displacy-arrowhead\" d=\"M1090.0,266.5 L1098.0,254.5 1082.0,254.5\" fill=\"currentColor\"/>\n",
|
| 314 |
"</g>\n",
|
| 315 |
"\n",
|
| 316 |
"<g class=\"displacy-arrow\">\n",
|
| 317 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-6\" stroke-width=\"2px\" d=\"M1295,264.5 C1295,177.0 1440.0,177.0 1440.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 318 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 319 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-6\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">det</textPath>\n",
|
| 320 |
" </text>\n",
|
| 321 |
+
" <path class=\"displacy-arrowhead\" d=\"M1295,266.5 L1287,254.5 1303,254.5\" fill=\"currentColor\"/>\n",
|
| 322 |
"</g>\n",
|
| 323 |
"\n",
|
| 324 |
"<g class=\"displacy-arrow\">\n",
|
| 325 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-7\" stroke-width=\"2px\" d=\"M1120,264.5 C1120,89.5 1445.0,89.5 1445.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 326 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 327 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-7\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">comp</textPath>\n",
|
| 328 |
" </text>\n",
|
| 329 |
+
" <path class=\"displacy-arrowhead\" d=\"M1445.0,266.5 L1453.0,254.5 1437.0,254.5\" fill=\"currentColor\"/>\n",
|
| 330 |
"</g>\n",
|
| 331 |
"\n",
|
| 332 |
"<g class=\"displacy-arrow\">\n",
|
| 333 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-8\" stroke-width=\"2px\" d=\"M1645,264.5 C1645,177.0 1790.0,177.0 1790.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 334 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 335 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-8\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">subj</textPath>\n",
|
| 336 |
" </text>\n",
|
| 337 |
+
" <path class=\"displacy-arrowhead\" d=\"M1645,266.5 L1637,254.5 1653,254.5\" fill=\"currentColor\"/>\n",
|
| 338 |
"</g>\n",
|
| 339 |
"\n",
|
| 340 |
"<g class=\"displacy-arrow\">\n",
|
| 341 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-9\" stroke-width=\"2px\" d=\"M1470,264.5 C1470,89.5 1795.0,89.5 1795.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 342 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 343 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-9\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">mod</textPath>\n",
|
| 344 |
" </text>\n",
|
| 345 |
+
" <path class=\"displacy-arrowhead\" d=\"M1795.0,266.5 L1803.0,254.5 1787.0,254.5\" fill=\"currentColor\"/>\n",
|
| 346 |
"</g>\n",
|
| 347 |
"\n",
|
| 348 |
"<g class=\"displacy-arrow\">\n",
|
| 349 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-10\" stroke-width=\"2px\" d=\"M1995,264.5 C1995,177.0 2140.0,177.0 2140.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 350 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 351 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-10\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">mod</textPath>\n",
|
| 352 |
" </text>\n",
|
| 353 |
+
" <path class=\"displacy-arrowhead\" d=\"M1995,266.5 L1987,254.5 2003,254.5\" fill=\"currentColor\"/>\n",
|
| 354 |
"</g>\n",
|
| 355 |
"\n",
|
| 356 |
"<g class=\"displacy-arrow\">\n",
|
| 357 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-11\" stroke-width=\"2px\" d=\"M1820,264.5 C1820,89.5 2145.0,89.5 2145.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 358 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 359 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-11\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">comp</textPath>\n",
|
| 360 |
" </text>\n",
|
| 361 |
+
" <path class=\"displacy-arrowhead\" d=\"M2145.0,266.5 L2153.0,254.5 2137.0,254.5\" fill=\"currentColor\"/>\n",
|
| 362 |
"</g>\n",
|
| 363 |
"\n",
|
| 364 |
"<g class=\"displacy-arrow\">\n",
|
| 365 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-12\" stroke-width=\"2px\" d=\"M2345,264.5 C2345,177.0 2490.0,177.0 2490.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 366 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 367 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-12\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">subj</textPath>\n",
|
| 368 |
" </text>\n",
|
| 369 |
+
" <path class=\"displacy-arrowhead\" d=\"M2345,266.5 L2337,254.5 2353,254.5\" fill=\"currentColor\"/>\n",
|
| 370 |
"</g>\n",
|
| 371 |
"\n",
|
| 372 |
"<g class=\"displacy-arrow\">\n",
|
| 373 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-13\" stroke-width=\"2px\" d=\"M595,264.5 C595,2.0 2500.0,2.0 2500.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 374 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 375 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-13\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">comp</textPath>\n",
|
| 376 |
" </text>\n",
|
| 377 |
+
" <path class=\"displacy-arrowhead\" d=\"M2500.0,266.5 L2508.0,254.5 2492.0,254.5\" fill=\"currentColor\"/>\n",
|
| 378 |
"</g>\n",
|
| 379 |
"\n",
|
| 380 |
"<g class=\"displacy-arrow\">\n",
|
| 381 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-14\" stroke-width=\"2px\" d=\"M2695,264.5 C2695,89.5 3020.0,89.5 3020.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 382 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 383 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-14\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">udep</textPath>\n",
|
| 384 |
" </text>\n",
|
| 385 |
+
" <path class=\"displacy-arrowhead\" d=\"M2695,266.5 L2687,254.5 2703,254.5\" fill=\"currentColor\"/>\n",
|
| 386 |
"</g>\n",
|
| 387 |
"\n",
|
| 388 |
"<g class=\"displacy-arrow\">\n",
|
| 389 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-15\" stroke-width=\"2px\" d=\"M2870,264.5 C2870,177.0 3015.0,177.0 3015.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 390 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 391 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-15\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">udep</textPath>\n",
|
| 392 |
" </text>\n",
|
| 393 |
+
" <path class=\"displacy-arrowhead\" d=\"M2870,266.5 L2862,254.5 2878,254.5\" fill=\"currentColor\"/>\n",
|
| 394 |
"</g>\n",
|
| 395 |
"\n",
|
| 396 |
"<g class=\"displacy-arrow\">\n",
|
| 397 |
+
" <path class=\"displacy-arc\" id=\"arrow-5232fcca8f23464a9783eb067e24fe32-0-16\" stroke-width=\"2px\" d=\"M2520,264.5 C2520,2.0 3025.0,2.0 3025.0,264.5\" fill=\"none\" stroke=\"currentColor\"/>\n",
|
| 398 |
" <text dy=\"1.25em\" style=\"font-size: 0.8em; letter-spacing: 1px\">\n",
|
| 399 |
+
" <textPath xlink:href=\"#arrow-5232fcca8f23464a9783eb067e24fe32-0-16\" class=\"displacy-label\" startOffset=\"50%\" side=\"left\" fill=\"currentColor\" text-anchor=\"middle\">comp</textPath>\n",
|
| 400 |
" </text>\n",
|
| 401 |
+
" <path class=\"displacy-arrowhead\" d=\"M3025.0,266.5 L3033.0,254.5 3017.0,254.5\" fill=\"currentColor\"/>\n",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 402 |
"</g>\n",
|
| 403 |
"</svg></span>"
|
| 404 |
],
|
|
|
|
| 411 |
}
|
| 412 |
],
|
| 413 |
"source": [
|
| 414 |
+
"doc = ssudify(nlp(\"In 1849 he began public readings of the story, which proved so successful he undertook 127 further performances.\"))\n",
|
| 415 |
"# Since this is an interactive Jupyter environment, we can use displacy.render here\n",
|
| 416 |
"displacy.render(doc, style='dep')"
|
| 417 |
]
|
| 418 |
},
|
| 419 |
{
|
| 420 |
"cell_type": "code",
|
| 421 |
+
"execution_count": 7,
|
| 422 |
"id": "daa1e8f5-2afe-4153-9469-a4c2af9b46b1",
|
| 423 |
"metadata": {},
|
| 424 |
"outputs": [
|
| 425 |
{
|
| 426 |
"data": {
|
| 427 |
"text/plain": [
|
| 428 |
+
"[(According to the studies, “People, 0),\n",
|
| 429 |
+
" (who danced more than once a week, 7),\n",
|
| 430 |
+
" (had, 0),\n",
|
| 431 |
+
" (a 76% lower risk of getting dementia, 8),\n",
|
| 432 |
+
" (compared to people who danced, 0),\n",
|
| 433 |
+
" (less frequently or, 3),\n",
|
| 434 |
+
" (didn’t dance at all,” Dr Pasricha noted., 0)]"
|
| 435 |
]
|
| 436 |
},
|
| 437 |
+
"execution_count": 7,
|
| 438 |
"metadata": {},
|
| 439 |
"output_type": "execute_result"
|
| 440 |
}
|
|
|
|
| 457 |
{
|
| 458 |
"cell_type": "code",
|
| 459 |
"execution_count": 9,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 460 |
"id": "8fb35fe3-68de-45b6-8053-8b390c8e8661",
|
| 461 |
"metadata": {},
|
| 462 |
"outputs": [],
|
|
|
|
| 466 |
},
|
| 467 |
{
|
| 468 |
"cell_type": "code",
|
| 469 |
+
"execution_count": 10,
|
| 470 |
"id": "9ce9c318-733d-4ee1-b5d7-39d98aa14697",
|
| 471 |
"metadata": {},
|
| 472 |
"outputs": [],
|
|
|
|
| 504 |
},
|
| 505 |
{
|
| 506 |
"cell_type": "code",
|
| 507 |
+
"execution_count": 11,
|
| 508 |
"id": "4cc68772-d26d-4cd1-be59-c23a5739c3d0",
|
| 509 |
"metadata": {},
|
| 510 |
"outputs": [
|
|
|
|
| 534 |
},
|
| 535 |
{
|
| 536 |
"cell_type": "code",
|
| 537 |
+
"execution_count": 12,
|
| 538 |
"id": "3e8c0a41-0d79-4715-ab2b-e5a2d0455569",
|
| 539 |
"metadata": {},
|
| 540 |
"outputs": [
|
|
|
|
| 553 |
"<IPython.core.display.HTML object>"
|
| 554 |
]
|
| 555 |
},
|
| 556 |
+
"execution_count": 12,
|
| 557 |
"metadata": {},
|
| 558 |
"output_type": "execute_result"
|
| 559 |
}
|
|
|
|
| 589 |
"name": "python",
|
| 590 |
"nbconvert_exporter": "python",
|
| 591 |
"pygments_lexer": "ipython3",
|
| 592 |
+
"version": "3.12.12"
|
| 593 |
}
|
| 594 |
},
|
| 595 |
"nbformat": 4,
|
main.py
CHANGED
|
@@ -12,6 +12,7 @@ relations = {
|
|
| 12 |
"udep": ["acl", "amod", "nmod", "poss", "nummod", "prep"],
|
| 13 |
}
|
| 14 |
|
|
|
|
| 15 |
def ssudify(doc):
|
| 16 |
for token in doc:
|
| 17 |
to_reverse = [token]
|
|
@@ -31,6 +32,11 @@ def ssudify(doc):
|
|
| 31 |
to_reverse[i-1].dep_ = "comp"
|
| 32 |
for token in doc:
|
| 33 |
if token.dep_ == "dep": token.dep_ = "unknown"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
for rel in relations.keys():
|
| 35 |
if token.dep_ in relations[rel]: token.dep_ = rel
|
| 36 |
for token in doc:
|
|
@@ -44,6 +50,10 @@ def ssudify(doc):
|
|
| 44 |
for child in [c for c in token.children if c.i < subject[0].i]:
|
| 45 |
if child.dep_ in ["comp", "udep"]:
|
| 46 |
child.dep_ = "mod"
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
return doc
|
| 48 |
|
| 49 |
|
|
|
|
| 12 |
"udep": ["acl", "amod", "nmod", "poss", "nummod", "prep"],
|
| 13 |
}
|
| 14 |
|
| 15 |
+
|
| 16 |
def ssudify(doc):
|
| 17 |
for token in doc:
|
| 18 |
to_reverse = [token]
|
|
|
|
| 32 |
to_reverse[i-1].dep_ = "comp"
|
| 33 |
for token in doc:
|
| 34 |
if token.dep_ == "dep": token.dep_ = "unknown"
|
| 35 |
+
if token.dep_ == "prep" and token.head.pos_ == "VERB" and token.i < token.head.i:
|
| 36 |
+
token.dep_ = "mod"
|
| 37 |
+
if token.dep_ == "ccomp" and any(sibling.dep_ in relations["comp"] + ["comp"]
|
| 38 |
+
for sibling in token.head.rights if sibling.i < token.i):
|
| 39 |
+
token.dep_ = "mod"
|
| 40 |
for rel in relations.keys():
|
| 41 |
if token.dep_ in relations[rel]: token.dep_ = rel
|
| 42 |
for token in doc:
|
|
|
|
| 50 |
for child in [c for c in token.children if c.i < subject[0].i]:
|
| 51 |
if child.dep_ in ["comp", "udep"]:
|
| 52 |
child.dep_ = "mod"
|
| 53 |
+
for token in doc:
|
| 54 |
+
if any(t.text == ";" for t in doc if (token.i < t.i < token.head.i or token.head.i < t.i < token.i) and token.pos_ != "PUNCT"):
|
| 55 |
+
token.head = token
|
| 56 |
+
token.dep_ = "root"
|
| 57 |
return doc
|
| 58 |
|
| 59 |
|