Uses 11335K of libraries. Click here for Pure Java version (6621L/42K).
!include once #1027304 // Eclipse Collections srecord noeq LCMerger_v2(LineCompReader lc1) { LineCompReader lc2; LineCompReader lcOut; // actually the same as lc1 now LCSortedPairIndex pairIndex; Map<S, Int> literalIndex; new IntBuffer lc2map; int newLiterals; // number of literals added compared to lc1 int newPairs; int nOriginalLiterals, nOriginalPairs; *(LineCompReader *lc1, LineCompReader *lc2) {} *(File ubcFile) { lc1 = LineCompReader(ubcFile); } void add(LineCompReader lc2) { //if (lc2 != null) fail("Can only run once for now"); this.lc2 = lc2; run(); } run { lcOut = lc1; nOriginalLiterals = l(lc1.literals); nOriginalPairs = l(lc1.pairs); newLiterals = newPairs = 0; printWithTime("Making literalIndex"); literalIndex = indexList(lc1.literals); // add lc2.literals to lc1.literals lc2map = new IntBuffer(l(lc2.literals) + l(lc2.pairs)); printWithTime("Merging literals"); for (int i = 0; i < l(lc2.literals); i++) { S c = lc2.literals.get(i); Int iLit = literalIndex.get(c); if (iLit == null) { iLit = addAndReturnIndex(lc1.literals, c); ++newLiterals; literalIndex.put(c, iLit); } lc2map.add(iLit); } // merge pairs printWithTime("Adjusting lc1 pairs"); for i to nOriginalPairs: { long p = lc1.pairs.get(i); lc1.pairs.set(i, twoIntsToLong(adjust1(firstIntFromLong(p)), adjust1(secondIntFromLong(p)))); } // Make index after adjustment printWithTime("Making pairIndex for " + nPairs(l(lc1.pairs))); pairIndex = new LCSortedPairIndex(lc1.pairs.toArray()); printWithTime("Merging pairs"); for (int i = 0; i < l(lc2.pairs); i++) { long p = lc2.pairs.get(i); long pAdjusted = twoIntsToLong(adjust2(firstIntFromLong(p)), adjust2(secondIntFromLong(p))); int iPair = pairIndex.get(pAdjusted); if (iPair < 0) { // new pair iPair = l(lc1.pairs); lc1.pairs.add(pAdjusted); ++newPairs; } lc2map.add(l(lcOut.literals)+iPair); } // copy files printWithTime("Merging files"); lc1.versions = (LinkedHashMap) mapValues(lc1.versions, enc -> lmap adjust1(enc)); for (S name, L<Int> encoding : lc2.versions) { if (lc1.versions.containsKey(name)) continue with print("Warning: Duplicate file name " + name); lc1.versions.put(name, lmap adjust2(encoding)); } printVars_str(+newLiterals, +newPairs); print("Synergy factor: " + doubleRatio(l(lc1.pairs), nOriginalPairs+l(lc2.pairs))); } // convert symbols from lc1 int adjust1(int i) { if (i >= nOriginalLiterals) if (i >= nOriginalLiterals+nOriginalPairs) ret i+newLiterals+newPairs; else ret i+newLiterals; ret i; } int adjust2(int i) { ret lc2map.get(i); } }
Began life as a copy of #1029327
download show line numbers debug dex old transpilations
Travelled to 7 computer(s): bhatertpkbcr, mqqgnosmbjvj, pyentgdyhuwx, pzhvpgtvlbxg, tvejysmllsmz, vouqrxazstgt, xrpafgyirdlv
No comments. add comment
Snippet ID: | #1029423 |
Snippet name: | LCMerger_v2 [with sorted pair index, uses less memory, dev.] |
Eternal ID of this version: | #1029423/8 |
Text MD5: | d45045d9ba943d419c037e1e2a8eaad0 |
Transpilation MD5: | ebb12ef1c62d91fe60dee5d8329ecb7c |
Author: | stefan |
Category: | javax |
Type: | JavaX fragment (include) |
Public (visible to everyone): | Yes |
Archived (hidden from active list): | No |
Created/modified: | 2020-08-05 19:23:44 |
Source code size: | 3013 bytes / 101 lines |
Pitched / IR pitched: | No / No |
Views / Downloads: | 236 / 526 |
Version history: | 7 change(s) |
Referenced in: | #1029424 - LCMerger_v3 [streaming lc2, OK, first one that actually works] #1034167 - Standard Classes + Interfaces (LIVE, continuation of #1003674) |