Not logged in.  Login/Logout/Register | List snippets | | Create snippet | Upload image | Upload data

99
LINES

< > BotCompany Repo | #1029357 // LCMerger [working backup]

JavaX fragment (include) [tags: use-pretranspiled]

Uses 11335K of libraries. Click here for Pure Java version (6398L/40K).

1  
!include once #1027304 // Eclipse Collections
2  
3  
srecord noeq LCMerger(LineCompReader lc1) {
4  
  LineCompReader lc2;
5  
  LineCompReader lcOut; // actually the same as lc1 now
6  
  LongIntHashMap pairIndex;
7  
  Map<S, Int> literalIndex;
8  
  IntIntHashMap lc2map; // TODO: Drop. This is actually contained in the pairIndex
9  
  int newLiterals; // number of literals added compared to lc1
10  
  int newPairs;
11  
  int nOriginalLiterals, nOriginalPairs;
12  
  
13  
  *(LineCompReader *lc1, LineCompReader *lc2) {}
14  
  
15  
  void add(LineCompReader lc2) {
16  
    this.lc2 = lc2;
17  
    run();
18  
  }
19  
20  
  run {
21  
    lcOut = lc1;
22  
    nOriginalLiterals = l(lc1.literals);
23  
    nOriginalPairs = l(lc1.pairs);
24  
    lc2map = new IntIntHashMap;
25  
    newLiterals = newPairs = 0;
26  
27  
    if (pairIndex != null) print("Have pairIndex");
28  
    else {
29  
      pairIndex = new LongIntHashMap;
30  
      printWithTime("Making pairIndex");
31  
      for (int i = 0; i < lc1.pairs.size(); i++)
32  
        pairIndex.put(lc1.pairs.get(i), i);
33  
    }
34  
    
35  
    printWithTime("Making literalIndex");
36  
    literalIndex = indexList(lc1.literals);
37  
  
38  
    // add lc2.literals to lc1.literals
39  
  
40  
    printWithTime("Merging literals");
41  
    for (int i = 0; i < l(lc2.literals); i++) {
42  
      S c = lc2.literals.get(i);
43  
      Int iLit = literalIndex.get(c);
44  
      if (iLit == null) {
45  
        iLit = addAndReturnIndex(lc1.literals, c);
46  
        ++newLiterals;
47  
        literalIndex.put(c, iLit);
48  
      }
49  
      lc2map.put(i, iLit);
50  
    }
51  
    
52  
    // merge pairs
53  
    
54  
    printWithTime("Merging pairs");
55  
    for i to nOriginalPairs: {
56  
      long p = lc1.pairs.get(i);
57  
      lc1.pairs.set(i, twoIntsToLong(adjust1(firstIntFromLong(p)), adjust1(secondIntFromLong(p))));
58  
    }
59  
    
60  
    for (int i = 0; i < l(lc2.pairs); i++) {
61  
      long p = lc2.pairs.get(i);
62  
      long pAdjusted = twoIntsToLong(adjust2(firstIntFromLong(p)), adjust2(secondIntFromLong(p)));
63  
      int iPair = pairIndex.getIfAbsent(pAdjusted, -1);
64  
      if (iPair < 0) { // new pair
65  
        iPair = l(lc1.pairs);
66  
        lc1.pairs.add(pAdjusted);
67  
        ++newPairs;
68  
        pairIndex.put(pAdjusted, iPair);
69  
      }
70  
      lc2map.put(l(lc2.literals)+i, l(lcOut.literals)+iPair);
71  
    }
72  
    
73  
    // copy files
74  
    
75  
    printWithTime("Merging files");
76  
    lc1.versions = (LinkedHashMap) mapValues(lc1.versions, enc -> lmap adjust1(enc));
77  
78  
    for (S name, L<Int> encoding : lc2.versions) {
79  
      if (lc1.versions.containsKey(name))
80  
        continue with print("Warning: Duplicate file name " + name);
81  
      lc1.versions.put(name, lmap adjust2(encoding));
82  
    }
83  
    
84  
    printVars_str(+newLiterals, +newPairs);
85  
    print("Synergy factor: " + doubleRatio(l(lc1.pairs), nOriginalPairs+l(lc2.pairs)));
86  
  }
87  
  
88  
  // convert symbols from lc1
89  
  int adjust1(int i) {
90  
    if (i >= nOriginalLiterals)
91  
      if (i >= nOriginalLiterals+nOriginalPairs)
92  
        ret i+newLiterals+newPairs;
93  
      else
94  
        ret i+newLiterals;
95  
    ret i;
96  
  }
97  
  
98  
  int adjust2(int i) { ret lc2map.getIfAbsent(i, i); }
99  
}

Author comment

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: #1029357
Snippet name: LCMerger [working backup]
Eternal ID of this version: #1029357/1
Text MD5: 24307a9d1e60eb76224cd3b4defec411
Transpilation MD5: ad3f8bc6b0afd5576cebff1ba7492711
Author: stefan
Category: javax
Type: JavaX fragment (include)
Public (visible to everyone): Yes
Archived (hidden from active list): No
Created/modified: 2020-08-02 13:43:54
Source code size: 3036 bytes / 99 lines
Pitched / IR pitched: No / No
Views / Downloads: 91 / 142
Referenced in: [show references]