vectra 0.5.3 → 0.5.5

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -14,6 +14,7 @@ export declare class TextSplitter {
14
14
  private recursiveSplit;
15
15
  private combineChunks;
16
16
  private containsAlphanumeric;
17
+ private splitBySpaces;
17
18
  private getSeparators;
18
19
  }
19
20
  //# sourceMappingURL=TextSplitter.d.ts.map
@@ -1 +1 @@
1
- {"version":3,"file":"TextSplitter.d.ts","sourceRoot":"","sources":["../src/TextSplitter.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,SAAS,CAAC;AAI/C,MAAM,WAAW,kBAAkB;IAC/B,UAAU,EAAE,MAAM,EAAE,CAAC;IACrB,cAAc,EAAE,OAAO,CAAC;IACxB,SAAS,EAAE,MAAM,CAAC;IAClB,YAAY,EAAE,MAAM,CAAC;IACrB,SAAS,EAAE,SAAS,CAAC;IACrB,OAAO,CAAC,EAAE,MAAM,CAAC;CACpB;AAED,qBAAa,YAAY;IACrB,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAqB;gBAE1B,MAAM,CAAC,EAAE,OAAO,CAAC,kBAAkB,CAAC;IA2BhD,KAAK,CAAC,IAAI,EAAE,MAAM,GAAG,SAAS,EAAE;IA4BvC,OAAO,CAAC,cAAc;IAoEtB,OAAO,CAAC,aAAa;IA6BrB,OAAO,CAAC,oBAAoB;IAS5B,OAAO,CAAC,aAAa;CA4WxB"}
1
+ {"version":3,"file":"TextSplitter.d.ts","sourceRoot":"","sources":["../src/TextSplitter.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,SAAS,CAAC;AAI/C,MAAM,WAAW,kBAAkB;IAC/B,UAAU,EAAE,MAAM,EAAE,CAAC;IACrB,cAAc,EAAE,OAAO,CAAC;IACxB,SAAS,EAAE,MAAM,CAAC;IAClB,YAAY,EAAE,MAAM,CAAC;IACrB,SAAS,EAAE,SAAS,CAAC;IACrB,OAAO,CAAC,EAAE,MAAM,CAAC;CACpB;AAED,qBAAa,YAAY;IACrB,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAqB;gBAE1B,MAAM,CAAC,EAAE,OAAO,CAAC,kBAAkB,CAAC;IA2BhD,KAAK,CAAC,IAAI,EAAE,MAAM,GAAG,SAAS,EAAE;IA4BvC,OAAO,CAAC,cAAc;IAoEtB,OAAO,CAAC,aAAa;IA6BrB,OAAO,CAAC,oBAAoB;IAS5B,OAAO,CAAC,aAAa;IAsBrB,OAAO,CAAC,aAAa;CA4WxB"}
@@ -64,7 +64,7 @@ class TextSplitter {
64
64
  if (separators.length > 0) {
65
65
  // Split by separator
66
66
  separator = separators[0];
67
- parts = text.split(separator);
67
+ parts = separator == ' ' ? this.splitBySpaces(text) : text.split(separator);
68
68
  }
69
69
  else {
70
70
  // Cut text in half
@@ -154,6 +154,28 @@ class TextSplitter {
154
154
  }
155
155
  return false;
156
156
  }
157
+ splitBySpaces(text) {
158
+ const parts = [];
159
+ const words = text.split(' ');
160
+ if (words.length > 0) {
161
+ let part = words[0];
162
+ for (let i = 1; i < words.length; i++) {
163
+ const nextWord = words[i];
164
+ if (this._config.tokenizer.encode(part + ' ' + nextWord).length <= this._config.chunkSize) {
165
+ part += ' ' + nextWord;
166
+ }
167
+ else {
168
+ parts.push(part);
169
+ part = nextWord;
170
+ }
171
+ }
172
+ parts.push(part);
173
+ }
174
+ else {
175
+ parts.push(text);
176
+ }
177
+ return parts;
178
+ }
157
179
  getSeparators(docType) {
158
180
  switch (docType !== null && docType !== void 0 ? docType : '') {
159
181
  case "cpp":
@@ -1 +1 @@
1
- {"version":3,"file":"TextSplitter.js","sourceRoot":"","sources":["../src/TextSplitter.ts"],"names":[],"mappings":";;;AAAA,mDAAgD;AAGhD,MAAM,kBAAkB,GAAG,gEAAgE,CAAC;AAW5F,MAAa,YAAY;IAGrB,YAAmB,MAAoC;QACnD,IAAI,CAAC,OAAO,GAAG,MAAM,CAAC,MAAM,CAAC;YACzB,cAAc,EAAE,KAAK;YACrB,SAAS,EAAE,GAAG;YACd,YAAY,EAAE,EAAE;SACG,EAAE,MAAM,CAAC,CAAC;QAEjC,iDAAiD;QACjD,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE;YACzB,IAAI,CAAC,OAAO,CAAC,SAAS,GAAG,IAAI,6BAAa,EAAE,CAAC;SAChD;QAED,8CAA8C;QAC9C,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,UAAU,IAAI,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,MAAM,KAAK,CAAC,EAAE;YAClE,IAAI,CAAC,OAAO,CAAC,UAAU,GAAG,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;SACtE;QAED,+BAA+B;QAC/B,IAAI,IAAI,CAAC,OAAO,CAAC,SAAS,GAAG,CAAC,EAAE;YAC5B,MAAM,IAAI,KAAK,CAAC,wBAAwB,CAAC,CAAC;SAC7C;aAAM,IAAI,IAAI,CAAC,OAAO,CAAC,YAAY,GAAG,CAAC,EAAE;YACtC,MAAM,IAAI,KAAK,CAAC,2BAA2B,CAAC,CAAC;SAChD;aAAM,IAAI,IAAI,CAAC,OAAO,CAAC,YAAY,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE;YAC3D,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;SACxD;IACL,CAAC;IAEM,KAAK,CAAC,IAAY;QACrB,mBAAmB;QACnB,MAAM,MAAM,GAAG,IAAI,CAAC,cAAc,CAAC,IAAI,EAAE,IAAI,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC;QAErE,MAAM,IAAI,GAAG,IAAI,CAAC;QAClB,SAAS,gBAAgB,CAAC,MAAiB;YACvC,IAAI,MAAM,IAAI,SAAS,EAAE;gBACrB,MAAM,GAAG,GAAG,MAAM,CAAC,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC;gBAClG,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;aAC/B;iBAAM;gBACH,OAAO,EAAE,CAAC;aACb;QACL,CAAC;QAED,iEAAiE;QACjE,IAAI,IAAI,CAAC,OAAO,CAAC,YAAY,GAAG,CAAC,EAAE;YAC/B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;gBACpC,MAAM,aAAa,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;gBACpC,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;gBACxB,MAAM,SAAS,GAAG,CAAC,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;gBACpE,KAAK,CAAC,YAAY,GAAG,gBAAgB,CAAC,aAAa,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC;gBAChF,KAAK,CAAC,UAAU,GAAG,gBAAgB,CAAC,SAAS,aAAT,SAAS,uBAAT,SAAS,CAAE,MAAM,CAAC,CAAC;aAC1D;SACJ;QAED,OAAO,MAAM,CAAC;IAClB,CAAC;IAEO,cAAc,CAAC,IAAY,EAAE,UAAoB,EAAE,QAAgB;QACvE,MAAM,MAAM,GAAgB,EAAE,CAAC;QAC/B,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE;YACjB,wBAAwB;YACxB,IAAI,KAAe,CAAC;YACpB,IAAI,SAAS,GAAG,EAAE,CAAC;YACnB,MAAM,cAAc,GAAG,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YACxE,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE;gBACvB,qBAAqB;gBACrB,SAAS,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;gBAC1B,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;aACjC;iBAAM;gBACH,mBAAmB;gBACnB,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;gBACzC,KAAK,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,EAAE,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC;aAC3D;YAED,qBAAqB;YACrB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;gBACnC,MAAM,SAAS,GAAG,CAAC,CAAC,KAAK,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;gBAE3C,4BAA4B;gBAC5B,IAAI,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;gBACrB,MAAM,MAAM,GAAG,CAAC,QAAQ,GAAG,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;gBACpF,IAAI,IAAI,CAAC,OAAO,CAAC,cAAc,IAAI,CAAC,SAAS,EAAE;oBAC3C,KAAK,IAAI,SAAS,CAAC;iBACtB;gBAED,6BAA6B;gBAC7B,IAAI,CAAC,IAAI,CAAC,oBAAoB,CAAC,KAAK,CAAC,EAAE;oBACnC,SAAS;iBACZ;gBAED,qDAAqD;gBACrD,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE;oBAC3C,qCAAqC;oBACrC,MAAM,SAAS,GAAG,IAAI,CAAC,cAAc,CAAC,KAAK,EAAE,cAAc,EAAE,QAAQ,CAAC,CAAC;oBACvE,MAAM,CAAC,IAAI,CAAC,GAAG,SAAS,CAAC,CAAC;iBAC7B;qBAAM;oBACH,oBAAoB;oBACpB,MAAM,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;oBACpD,IAAI,MAAM,CAAC,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE;wBACxC,qCAAqC;wBACrC,MAAM,SAAS,GAAG,IAAI,CAAC,cAAc,CAAC,KAAK,EAAE,cAAc,EAAE,QAAQ,CAAC,CAAC;wBACvE,MAAM,CAAC,IAAI,CAAC,GAAG,SAAS,CAAC,CAAC;qBAC7B;yBAAM;wBACH,yBAAyB;wBACzB,MAAM,CAAC,IAAI,CAAC;4BACR,IAAI,EAAE,KAAK;4BACX,MAAM,EAAE,MAAM;4BACd,QAAQ,EAAE,QAAQ;4BAClB,MAAM,EAAE,MAAM;4BACd,YAAY,EAAE,EAAE;4BAChB,UAAU,EAAE,EAAE;yBACjB,CAAC,CAAC;qBACN;iBAEJ;gBAGD,kBAAkB;gBAClB,QAAQ,GAAG,MAAM,GAAG,CAAC,CAAC;aACzB;SACJ;QAED,OAAO,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,CAAC;IACtC,CAAC;IAEO,aAAa,CAAC,MAAmB;QACrC,MAAM,cAAc,GAAgB,EAAE,CAAC;QACvC,IAAI,YAAiC,CAAC;QACtC,IAAI,aAAa,GAAG,CAAC,CAAC;QACtB,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC;QACzD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;YACpC,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;YACxB,IAAI,YAAY,EAAE;gBACd,MAAM,MAAM,GAAG,YAAY,CAAC,MAAM,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC;gBAChE,IAAI,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE;oBACjC,cAAc,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;oBAClC,YAAY,GAAG,KAAK,CAAC;oBACrB,aAAa,GAAG,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC;iBACvC;qBAAM;oBACH,YAAY,CAAC,IAAI,IAAI,SAAS,GAAG,KAAK,CAAC,IAAI,CAAC;oBAC5C,YAAY,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC;oBAC1C,aAAa,IAAI,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC;iBACxC;aACJ;iBAAM;gBACH,YAAY,GAAG,KAAK,CAAC;gBACrB,aAAa,GAAG,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC;aACvC;SACJ;QACD,IAAI,YAAY,EAAE;YACd,cAAc,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;SACrC;QACD,OAAO,cAAc,CAAC;IAC1B,CAAC;IAEO,oBAAoB,CAAC,IAAY;QACrC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;YAClC,IAAI,kBAAkB,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE;gBACtC,OAAO,IAAI,CAAC;aACf;SACJ;QACD,OAAO,KAAK,CAAC;IACjB,CAAC;IAEO,aAAa,CAAC,OAAgB;QAClC,QAAQ,OAAO,aAAP,OAAO,cAAP,OAAO,GAAI,EAAE,EAAE;YACnB,KAAK,KAAK;gBACN,OAAO;oBACH,gCAAgC;oBAChC,UAAU;oBACV,mCAAmC;oBACnC,SAAS;oBACT,QAAQ;oBACR,UAAU;oBACV,WAAW;oBACX,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,UAAU;oBACV,WAAW;oBACX,SAAS;oBACT,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,IAAI;gBACL,OAAO;oBACH,mCAAmC;oBACnC,SAAS;oBACT,QAAQ;oBACR,UAAU;oBACV,SAAS;oBACT,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,WAAW;oBACX,SAAS;oBACT,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,MAAM,CAAC;YACZ,KAAK,IAAI,CAAC;YACV,KAAK,QAAQ,CAAC;YACd,KAAK,IAAI,CAAC;YACV,KAAK,IAAI,CAAC;YACV,KAAK,KAAK,CAAC;YACX,KAAK,YAAY;gBACb,OAAO;oBACH,sBAAsB;oBACtB,eAAe;oBACf,eAAe;oBACf,gBAAgB;oBAChB,gCAAgC;oBAChC,UAAU;oBACV,iCAAiC;oBACjC,WAAW;oBACX,cAAc;oBACd,YAAY;oBACZ,WAAW;oBACX,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,UAAU;oBACV,WAAW;oBACX,SAAS;oBACT,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,IAAI,CAAC;YACV,KAAK,KAAK,CAAC;YACX,KAAK,YAAY;gBACb,OAAO;oBACH,sBAAsB;oBACtB,eAAe;oBACf,eAAe;oBACf,gBAAgB;oBAChB,gCAAgC;oBAChC,UAAU;oBACV,mCAAmC;oBACnC,aAAa;oBACb,UAAU;oBACV,QAAQ;oBACR,QAAQ;oBACR,UAAU;oBACV,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,UAAU;oBACV,WAAW;oBACX,SAAS;oBACT,YAAY;oBACZ,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,KAAK;gBACN,OAAO;oBACH,mCAAmC;oBACnC,aAAa;oBACb,gCAAgC;oBAChC,UAAU;oBACV,sCAAsC;oBACtC,OAAO;oBACP,YAAY;oBACZ,UAAU;oBACV,OAAO;oBACP,WAAW;oBACX,SAAS;oBACT,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,OAAO;gBACR,OAAO;oBACH,kCAAkC;oBAClC,YAAY;oBACZ,kCAAkC;oBAClC,YAAY;oBACZ,+BAA+B;oBAC/B,SAAS;oBACT,iCAAiC;oBACjC,WAAW;oBACX,gCAAgC;oBAChC,WAAW;oBACX,kCAAkC;oBAClC,WAAW;oBACX,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,QAAQ,CAAC;YACd,KAAK,IAAI;gBACL,OAAO;oBACH,8CAA8C;oBAC9C,UAAU;oBACV,QAAQ;oBACR,UAAU;oBACV,wCAAwC;oBACxC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,KAAK;gBACN,OAAO;oBACH,6BAA6B;oBAC7B,SAAS;oBACT,SAAS;oBACT,SAAS;oBACT,gCAAgC;oBAChC,OAAO;oBACP,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,MAAM;gBACP,OAAO;oBACH,iCAAiC;oBACjC,QAAQ;oBACR,UAAU;oBACV,sCAAsC;oBACtC,OAAO;oBACP,WAAW;oBACX,UAAU;oBACV,QAAQ;oBACR,OAAO;oBACP,UAAU;oBACV,WAAW;oBACX,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,MAAM;gBACP,OAAO;oBACH,mCAAmC;oBACnC,OAAO;oBACP,UAAU;oBACV,QAAQ;oBACR,sCAAsC;oBACtC,OAAO;oBACP,UAAU;oBACV,QAAQ;oBACR,SAAS;oBACT,UAAU;oBACV,UAAU;oBACV,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,OAAO;gBACR,OAAO;oBACH,gCAAgC;oBAChC,UAAU;oBACV,WAAW;oBACX,iCAAiC;oBACjC,QAAQ;oBACR,QAAQ;oBACR,QAAQ;oBACR,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,UAAU;oBACV,UAAU;oBACV,SAAS;oBACT,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,OAAO;gBACR,OAAO;oBACH,mCAAmC;oBACnC,SAAS;oBACT,gCAAgC;oBAChC,UAAU;oBACV,WAAW;oBACX,SAAS;oBACT,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,UAAU;oBACV,OAAO;oBACP,WAAW;oBACX,SAAS;oBACT,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,IAAI,CAAC;YACV,KAAK,UAAU;gBACX,OAAO;oBACH,sEAAsE;oBACtE,OAAO;oBACP,QAAQ;oBACR,SAAS;oBACT,UAAU;oBACV,WAAW;oBACX,uEAAuE;oBACvE,kBAAkB;oBAClB,kBAAkB;oBAClB,oBAAoB;oBACpB,SAAS;oBACT,mBAAmB;oBACnB,aAAa;oBACb,aAAa;oBACb,aAAa;oBACb,kEAAkE;oBAClE,kEAAkE;oBAClE,gBAAgB;oBAChB,SAAS;oBACT,UAAU;oBACV,UAAU;oBACV,UAAU;oBACV,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,OAAO;gBACR,OAAO;oBACH,2CAA2C;oBAC3C,cAAc;oBACd,cAAc;oBACd,iBAAiB;oBACjB,oBAAoB;oBAEpB,4BAA4B;oBAC5B,sBAAsB;oBACtB,oBAAoB;oBACpB,wBAAwB;oBACxB,iBAAiB;oBACjB,kBAAkB;oBAClB,sBAAsB;oBACtB,kBAAkB;oBAClB,qBAAqB;oBAErB,iCAAiC;oBACjC,kBAAkB;oBAClB,IAAI;oBACJ,GAAG;oBAEH,wCAAwC;oBACxC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,MAAM;gBACP,OAAO;oBACH,sCAAsC;oBACtC,QAAQ;oBACR,OAAO;oBACP,KAAK;oBACL,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,QAAQ;oBACR,SAAS;oBACT,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,UAAU;oBACV,UAAU;oBACV,OAAO;oBACP,OAAO;oBACP,QAAQ;oBACR,SAAS;oBACT,UAAU;oBACV,QAAQ;oBACR,SAAS;oBACT,uBAAuB;oBACvB,GAAG;iBACN,CAAC;YACN,KAAK,KAAK;gBACN,OAAO;oBACH,gDAAgD;oBAChD,WAAW;oBACX,UAAU;oBACV,mCAAmC;oBACnC,aAAa;oBACb,cAAc;oBACd,YAAY;oBACZ,iCAAiC;oBACjC,gBAAgB;oBAChB,SAAS;oBACT,aAAa;oBACb,UAAU;oBACV,aAAa;oBACb,UAAU;oBACV,WAAW;oBACX,SAAS;oBACT,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,UAAU;oBACV,aAAa;oBACb,aAAa;oBACb,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN;gBACI,OAAO;oBACH,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;oBACH,EAAE;iBACL,CAAC;SACT;IACL,CAAC;CACJ;AAhhBD,oCAghBC"}
1
+ {"version":3,"file":"TextSplitter.js","sourceRoot":"","sources":["../src/TextSplitter.ts"],"names":[],"mappings":";;;AAAA,mDAAgD;AAGhD,MAAM,kBAAkB,GAAG,gEAAgE,CAAC;AAW5F,MAAa,YAAY;IAGrB,YAAmB,MAAoC;QACnD,IAAI,CAAC,OAAO,GAAG,MAAM,CAAC,MAAM,CAAC;YACzB,cAAc,EAAE,KAAK;YACrB,SAAS,EAAE,GAAG;YACd,YAAY,EAAE,EAAE;SACG,EAAE,MAAM,CAAC,CAAC;QAEjC,iDAAiD;QACjD,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE;YACzB,IAAI,CAAC,OAAO,CAAC,SAAS,GAAG,IAAI,6BAAa,EAAE,CAAC;SAChD;QAED,8CAA8C;QAC9C,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,UAAU,IAAI,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,MAAM,KAAK,CAAC,EAAE;YAClE,IAAI,CAAC,OAAO,CAAC,UAAU,GAAG,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;SACtE;QAED,+BAA+B;QAC/B,IAAI,IAAI,CAAC,OAAO,CAAC,SAAS,GAAG,CAAC,EAAE;YAC5B,MAAM,IAAI,KAAK,CAAC,wBAAwB,CAAC,CAAC;SAC7C;aAAM,IAAI,IAAI,CAAC,OAAO,CAAC,YAAY,GAAG,CAAC,EAAE;YACtC,MAAM,IAAI,KAAK,CAAC,2BAA2B,CAAC,CAAC;SAChD;aAAM,IAAI,IAAI,CAAC,OAAO,CAAC,YAAY,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE;YAC3D,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;SACxD;IACL,CAAC;IAEM,KAAK,CAAC,IAAY;QACrB,mBAAmB;QACnB,MAAM,MAAM,GAAG,IAAI,CAAC,cAAc,CAAC,IAAI,EAAE,IAAI,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC;QAErE,MAAM,IAAI,GAAG,IAAI,CAAC;QAClB,SAAS,gBAAgB,CAAC,MAAiB;YACvC,IAAI,MAAM,IAAI,SAAS,EAAE;gBACrB,MAAM,GAAG,GAAG,MAAM,CAAC,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC;gBAClG,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;aAC/B;iBAAM;gBACH,OAAO,EAAE,CAAC;aACb;QACL,CAAC;QAED,iEAAiE;QACjE,IAAI,IAAI,CAAC,OAAO,CAAC,YAAY,GAAG,CAAC,EAAE;YAC/B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;gBACpC,MAAM,aAAa,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;gBACpC,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;gBACxB,MAAM,SAAS,GAAG,CAAC,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;gBACpE,KAAK,CAAC,YAAY,GAAG,gBAAgB,CAAC,aAAa,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC;gBAChF,KAAK,CAAC,UAAU,GAAG,gBAAgB,CAAC,SAAS,aAAT,SAAS,uBAAT,SAAS,CAAE,MAAM,CAAC,CAAC;aAC1D;SACJ;QAED,OAAO,MAAM,CAAC;IAClB,CAAC;IAEO,cAAc,CAAC,IAAY,EAAE,UAAoB,EAAE,QAAgB;QACvE,MAAM,MAAM,GAAgB,EAAE,CAAC;QAC/B,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE;YACjB,wBAAwB;YACxB,IAAI,KAAe,CAAC;YACpB,IAAI,SAAS,GAAG,EAAE,CAAC;YACnB,MAAM,cAAc,GAAG,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YACxE,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE;gBACvB,qBAAqB;gBACrB,SAAS,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;gBAC1B,KAAK,GAAG,SAAS,IAAI,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;aAC/E;iBAAM;gBACH,mBAAmB;gBACnB,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;gBACzC,KAAK,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,EAAE,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC;aAC3D;YAED,qBAAqB;YACrB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;gBACnC,MAAM,SAAS,GAAG,CAAC,CAAC,KAAK,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;gBAE3C,4BAA4B;gBAC5B,IAAI,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;gBACrB,MAAM,MAAM,GAAG,CAAC,QAAQ,GAAG,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;gBACpF,IAAI,IAAI,CAAC,OAAO,CAAC,cAAc,IAAI,CAAC,SAAS,EAAE;oBAC3C,KAAK,IAAI,SAAS,CAAC;iBACtB;gBAED,6BAA6B;gBAC7B,IAAI,CAAC,IAAI,CAAC,oBAAoB,CAAC,KAAK,CAAC,EAAE;oBACnC,SAAS;iBACZ;gBAED,qDAAqD;gBACrD,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE;oBAC3C,qCAAqC;oBACrC,MAAM,SAAS,GAAG,IAAI,CAAC,cAAc,CAAC,KAAK,EAAE,cAAc,EAAE,QAAQ,CAAC,CAAC;oBACvE,MAAM,CAAC,IAAI,CAAC,GAAG,SAAS,CAAC,CAAC;iBAC7B;qBAAM;oBACH,oBAAoB;oBACpB,MAAM,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;oBACpD,IAAI,MAAM,CAAC,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE;wBACxC,qCAAqC;wBACrC,MAAM,SAAS,GAAG,IAAI,CAAC,cAAc,CAAC,KAAK,EAAE,cAAc,EAAE,QAAQ,CAAC,CAAC;wBACvE,MAAM,CAAC,IAAI,CAAC,GAAG,SAAS,CAAC,CAAC;qBAC7B;yBAAM;wBACH,yBAAyB;wBACzB,MAAM,CAAC,IAAI,CAAC;4BACR,IAAI,EAAE,KAAK;4BACX,MAAM,EAAE,MAAM;4BACd,QAAQ,EAAE,QAAQ;4BAClB,MAAM,EAAE,MAAM;4BACd,YAAY,EAAE,EAAE;4BAChB,UAAU,EAAE,EAAE;yBACjB,CAAC,CAAC;qBACN;iBAEJ;gBAGD,kBAAkB;gBAClB,QAAQ,GAAG,MAAM,GAAG,CAAC,CAAC;aACzB;SACJ;QAED,OAAO,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,CAAC;IACtC,CAAC;IAEO,aAAa,CAAC,MAAmB;QACrC,MAAM,cAAc,GAAgB,EAAE,CAAC;QACvC,IAAI,YAAiC,CAAC;QACtC,IAAI,aAAa,GAAG,CAAC,CAAC;QACtB,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC;QACzD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;YACpC,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;YACxB,IAAI,YAAY,EAAE;gBACd,MAAM,MAAM,GAAG,YAAY,CAAC,MAAM,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC;gBAChE,IAAI,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE;oBACjC,cAAc,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;oBAClC,YAAY,GAAG,KAAK,CAAC;oBACrB,aAAa,GAAG,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC;iBACvC;qBAAM;oBACH,YAAY,CAAC,IAAI,IAAI,SAAS,GAAG,KAAK,CAAC,IAAI,CAAC;oBAC5C,YAAY,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC;oBAC1C,aAAa,IAAI,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC;iBACxC;aACJ;iBAAM;gBACH,YAAY,GAAG,KAAK,CAAC;gBACrB,aAAa,GAAG,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC;aACvC;SACJ;QACD,IAAI,YAAY,EAAE;YACd,cAAc,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;SACrC;QACD,OAAO,cAAc,CAAC;IAC1B,CAAC;IAEO,oBAAoB,CAAC,IAAY;QACrC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;YAClC,IAAI,kBAAkB,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE;gBACtC,OAAO,IAAI,CAAC;aACf;SACJ;QACD,OAAO,KAAK,CAAC;IACjB,CAAC;IAEO,aAAa,CAAC,IAAY;QAC9B,MAAM,KAAK,GAAa,EAAE,CAAC;QAC3B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QAC9B,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE;YAClB,IAAI,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;YACpB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;gBACnC,MAAM,QAAQ,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;gBAC1B,IAAI,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,MAAM,CAAC,IAAI,GAAG,GAAG,GAAG,QAAQ,CAAC,CAAC,MAAM,IAAI,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE;oBACvF,IAAI,IAAI,GAAG,GAAG,QAAQ,CAAC;iBAC1B;qBAAM;oBACH,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;oBACjB,IAAI,GAAG,QAAQ,CAAC;iBACnB;aACJ;YACD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;SACpB;aAAM;YACH,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;SACpB;QAED,OAAO,KAAK,CAAC;IACjB,CAAC;IAEO,aAAa,CAAC,OAAgB;QAClC,QAAQ,OAAO,aAAP,OAAO,cAAP,OAAO,GAAI,EAAE,EAAE;YACnB,KAAK,KAAK;gBACN,OAAO;oBACH,gCAAgC;oBAChC,UAAU;oBACV,mCAAmC;oBACnC,SAAS;oBACT,QAAQ;oBACR,UAAU;oBACV,WAAW;oBACX,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,UAAU;oBACV,WAAW;oBACX,SAAS;oBACT,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,IAAI;gBACL,OAAO;oBACH,mCAAmC;oBACnC,SAAS;oBACT,QAAQ;oBACR,UAAU;oBACV,SAAS;oBACT,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,WAAW;oBACX,SAAS;oBACT,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,MAAM,CAAC;YACZ,KAAK,IAAI,CAAC;YACV,KAAK,QAAQ,CAAC;YACd,KAAK,IAAI,CAAC;YACV,KAAK,IAAI,CAAC;YACV,KAAK,KAAK,CAAC;YACX,KAAK,YAAY;gBACb,OAAO;oBACH,sBAAsB;oBACtB,eAAe;oBACf,eAAe;oBACf,gBAAgB;oBAChB,gCAAgC;oBAChC,UAAU;oBACV,iCAAiC;oBACjC,WAAW;oBACX,cAAc;oBACd,YAAY;oBACZ,WAAW;oBACX,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,UAAU;oBACV,WAAW;oBACX,SAAS;oBACT,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,IAAI,CAAC;YACV,KAAK,KAAK,CAAC;YACX,KAAK,YAAY;gBACb,OAAO;oBACH,sBAAsB;oBACtB,eAAe;oBACf,eAAe;oBACf,gBAAgB;oBAChB,gCAAgC;oBAChC,UAAU;oBACV,mCAAmC;oBACnC,aAAa;oBACb,UAAU;oBACV,QAAQ;oBACR,QAAQ;oBACR,UAAU;oBACV,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,UAAU;oBACV,WAAW;oBACX,SAAS;oBACT,YAAY;oBACZ,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,KAAK;gBACN,OAAO;oBACH,mCAAmC;oBACnC,aAAa;oBACb,gCAAgC;oBAChC,UAAU;oBACV,sCAAsC;oBACtC,OAAO;oBACP,YAAY;oBACZ,UAAU;oBACV,OAAO;oBACP,WAAW;oBACX,SAAS;oBACT,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,OAAO;gBACR,OAAO;oBACH,kCAAkC;oBAClC,YAAY;oBACZ,kCAAkC;oBAClC,YAAY;oBACZ,+BAA+B;oBAC/B,SAAS;oBACT,iCAAiC;oBACjC,WAAW;oBACX,gCAAgC;oBAChC,WAAW;oBACX,kCAAkC;oBAClC,WAAW;oBACX,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,QAAQ,CAAC;YACd,KAAK,IAAI;gBACL,OAAO;oBACH,8CAA8C;oBAC9C,UAAU;oBACV,QAAQ;oBACR,UAAU;oBACV,wCAAwC;oBACxC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,KAAK;gBACN,OAAO;oBACH,6BAA6B;oBAC7B,SAAS;oBACT,SAAS;oBACT,SAAS;oBACT,gCAAgC;oBAChC,OAAO;oBACP,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,MAAM;gBACP,OAAO;oBACH,iCAAiC;oBACjC,QAAQ;oBACR,UAAU;oBACV,sCAAsC;oBACtC,OAAO;oBACP,WAAW;oBACX,UAAU;oBACV,QAAQ;oBACR,OAAO;oBACP,UAAU;oBACV,WAAW;oBACX,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,MAAM;gBACP,OAAO;oBACH,mCAAmC;oBACnC,OAAO;oBACP,UAAU;oBACV,QAAQ;oBACR,sCAAsC;oBACtC,OAAO;oBACP,UAAU;oBACV,QAAQ;oBACR,SAAS;oBACT,UAAU;oBACV,UAAU;oBACV,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,OAAO;gBACR,OAAO;oBACH,gCAAgC;oBAChC,UAAU;oBACV,WAAW;oBACX,iCAAiC;oBACjC,QAAQ;oBACR,QAAQ;oBACR,QAAQ;oBACR,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,UAAU;oBACV,UAAU;oBACV,SAAS;oBACT,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,OAAO;gBACR,OAAO;oBACH,mCAAmC;oBACnC,SAAS;oBACT,gCAAgC;oBAChC,UAAU;oBACV,WAAW;oBACX,SAAS;oBACT,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,UAAU;oBACV,OAAO;oBACP,WAAW;oBACX,SAAS;oBACT,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,IAAI,CAAC;YACV,KAAK,UAAU;gBACX,OAAO;oBACH,sEAAsE;oBACtE,OAAO;oBACP,QAAQ;oBACR,SAAS;oBACT,UAAU;oBACV,WAAW;oBACX,uEAAuE;oBACvE,kBAAkB;oBAClB,kBAAkB;oBAClB,oBAAoB;oBACpB,SAAS;oBACT,mBAAmB;oBACnB,aAAa;oBACb,aAAa;oBACb,aAAa;oBACb,kEAAkE;oBAClE,kEAAkE;oBAClE,gBAAgB;oBAChB,SAAS;oBACT,UAAU;oBACV,UAAU;oBACV,UAAU;oBACV,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,OAAO;gBACR,OAAO;oBACH,2CAA2C;oBAC3C,cAAc;oBACd,cAAc;oBACd,iBAAiB;oBACjB,oBAAoB;oBAEpB,4BAA4B;oBAC5B,sBAAsB;oBACtB,oBAAoB;oBACpB,wBAAwB;oBACxB,iBAAiB;oBACjB,kBAAkB;oBAClB,sBAAsB;oBACtB,kBAAkB;oBAClB,qBAAqB;oBAErB,iCAAiC;oBACjC,kBAAkB;oBAClB,IAAI;oBACJ,GAAG;oBAEH,wCAAwC;oBACxC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN,KAAK,MAAM;gBACP,OAAO;oBACH,sCAAsC;oBACtC,QAAQ;oBACR,OAAO;oBACP,KAAK;oBACL,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,QAAQ;oBACR,SAAS;oBACT,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,MAAM;oBACN,UAAU;oBACV,UAAU;oBACV,OAAO;oBACP,OAAO;oBACP,QAAQ;oBACR,SAAS;oBACT,UAAU;oBACV,QAAQ;oBACR,SAAS;oBACT,uBAAuB;oBACvB,GAAG;iBACN,CAAC;YACN,KAAK,KAAK;gBACN,OAAO;oBACH,gDAAgD;oBAChD,WAAW;oBACX,UAAU;oBACV,mCAAmC;oBACnC,aAAa;oBACb,cAAc;oBACd,YAAY;oBACZ,iCAAiC;oBACjC,gBAAgB;oBAChB,SAAS;oBACT,aAAa;oBACb,UAAU;oBACV,aAAa;oBACb,UAAU;oBACV,WAAW;oBACX,SAAS;oBACT,sCAAsC;oBACtC,OAAO;oBACP,QAAQ;oBACR,UAAU;oBACV,aAAa;oBACb,aAAa;oBACb,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;iBACN,CAAC;YACN;gBACI,OAAO;oBACH,oCAAoC;oBACpC,MAAM;oBACN,IAAI;oBACJ,GAAG;oBACH,EAAE;iBACL,CAAC;SACT;IACL,CAAC;CACJ;AAtiBD,oCAsiBC"}
package/lib/vectra-cli.js CHANGED
@@ -118,7 +118,7 @@ function run() {
118
118
  });
119
119
  // Get list of url's
120
120
  const uris = yield getItemList(args.uri, args.list, 'web page');
121
- // Fetch web pages
121
+ // Fetch documents
122
122
  const fileFetcher = new FileFetcher_1.FileFetcher();
123
123
  const webFetcher = args.cookie ? new WebFetcher_1.WebFetcher({ headers: { "cookie": args.cookie } }) : new WebFetcher_1.WebFetcher();
124
124
  for (const path of uris) {
package/package.json CHANGED
@@ -2,7 +2,7 @@
2
2
  "name": "vectra",
3
3
  "author": "Steven Ickman",
4
4
  "description": "A vector database that uses the local file system for storage.",
5
- "version": "0.5.3",
5
+ "version": "0.5.5",
6
6
  "license": "MIT",
7
7
  "keywords": [
8
8
  "gpt"
@@ -80,7 +80,7 @@ export class TextSplitter {
80
80
  if (separators.length > 0) {
81
81
  // Split by separator
82
82
  separator = separators[0];
83
- parts = text.split(separator);
83
+ parts = separator == ' ' ? this.splitBySpaces(text) : text.split(separator);
84
84
  } else {
85
85
  // Cut text in half
86
86
  const half = Math.floor(text.length / 2);
@@ -176,6 +176,28 @@ export class TextSplitter {
176
176
  return false;
177
177
  }
178
178
 
179
+ private splitBySpaces(text: string): string[] {
180
+ const parts: string[] = [];
181
+ const words = text.split(' ');
182
+ if (words.length > 0) {
183
+ let part = words[0];
184
+ for (let i = 1; i < words.length; i++) {
185
+ const nextWord = words[i];
186
+ if (this._config.tokenizer.encode(part + ' ' + nextWord).length <= this._config.chunkSize) {
187
+ part += ' ' + nextWord;
188
+ } else {
189
+ parts.push(part);
190
+ part = nextWord;
191
+ }
192
+ }
193
+ parts.push(part);
194
+ } else {
195
+ parts.push(text);
196
+ }
197
+
198
+ return parts;
199
+ }
200
+
179
201
  private getSeparators(docType?: string): string[] {
180
202
  switch (docType ?? '') {
181
203
  case "cpp":
package/src/vectra-cli.ts CHANGED
@@ -82,7 +82,7 @@ export async function run() {
82
82
  // Get list of url's
83
83
  const uris = await getItemList(args.uri as string[], args.list as string, 'web page');
84
84
 
85
- // Fetch web pages
85
+ // Fetch documents
86
86
  const fileFetcher = new FileFetcher();
87
87
  const webFetcher = args.cookie ? new WebFetcher({ headers: { "cookie": args.cookie }}) : new WebFetcher();
88
88
  for (const path of uris) {