website/archive/binaries/mac/src/muscle/phyfromfile.cpp

   1 #include "muscle.h"\r
   2 #include "tree.h"\r
   3 #include "textfile.h"\r
   4 \r
   5 #define TRACE 0\r
   6 \r
   7 // Tokens in Newick files are:\r
   8 //              ( ) : , ;\r
   9 //              string\r
  10 //              'string'\r
  11 //              "string"\r
  12 //              [ comment ]\r
  13 //\r
  14 // We can't safely distinguish between identifiers and floating point\r
  15 // numbers at the lexical level (because identifiers may be numeric,\r
  16 // or start with digits), so both edge lengths and identifiers are\r
  17 // returned as strings.\r
  18 \r
  19 const char *Tree::NTTStr(NEWICK_TOKEN_TYPE NTT) const\r
  20         {\r
  21         switch (NTT)\r
  22                 {\r
  23 #define c(x)    case NTT_##x: return #x;\r
  24         c(Unknown)\r
  25         c(Lparen)\r
  26         c(Rparen)\r
  27         c(Colon)\r
  28         c(Comma)\r
  29         c(Semicolon)\r
  30         c(String)\r
  31         c(SingleQuotedString)\r
  32         c(DoubleQuotedString)\r
  33         c(Comment)\r
  34 #undef c\r
  35                 }\r
  36         return "??";\r
  37         }\r
  38 \r
  39 NEWICK_TOKEN_TYPE Tree::GetToken(TextFile &File, char szToken[], unsigned uBytes) const\r
  40         {\r
  41 // Skip leading white space\r
  42         File.SkipWhite();\r
  43 \r
  44         char c;\r
  45         File.GetCharX(c);\r
  46 \r
  47 // In case a single-character token\r
  48         szToken[0] = c;\r
  49         szToken[1] = 0;\r
  50 \r
  51         unsigned uBytesCopied = 0;\r
  52         NEWICK_TOKEN_TYPE TT;\r
  53         switch (c)\r
  54                 {\r
  55         case '(':\r
  56                 return NTT_Lparen;\r
  57 \r
  58         case ')':\r
  59                 return NTT_Rparen;\r
  60 \r
  61         case ':':\r
  62                 return NTT_Colon;\r
  63 \r
  64         case ';':\r
  65                 return NTT_Semicolon;\r
  66 \r
  67         case ',':\r
  68                 return NTT_Comma;\r
  69 \r
  70         case '\'':\r
  71                 TT = NTT_SingleQuotedString;\r
  72                 File.GetCharX(c);\r
  73                 break;\r
  74 \r
  75         case '"':\r
  76                 TT = NTT_DoubleQuotedString;\r
  77                 File.GetCharX(c);\r
  78                 break;\r
  79 \r
  80         case '[':\r
  81                 TT = NTT_Comment;\r
  82                 break;\r
  83 \r
  84         default:\r
  85                 TT = NTT_String;\r
  86                 break;\r
  87                 }\r
  88 \r
  89         for (;;)\r
  90                 {\r
  91                 if (TT != NTT_Comment)\r
  92                         {\r
  93                         if (uBytesCopied < uBytes - 2)\r
  94                                 {\r
  95                                 szToken[uBytesCopied++] = c;\r
  96                                 szToken[uBytesCopied] = 0;\r
  97                                 }\r
  98                         else\r
  99                                 Quit("Tree::GetToken: input buffer too small, token so far='%s'", szToken);\r
 100                         }\r
 101                 bool bEof = File.GetChar(c);\r
 102                 if (bEof)\r
 103                         return TT;\r
 104 \r
 105                 switch (TT)\r
 106                         {\r
 107                 case NTT_String:\r
 108                         if (0 != strchr("():;,", c))\r
 109                                 {\r
 110                                 File.PushBack(c);\r
 111                                 return NTT_String;\r
 112                                 }\r
 113                         if (isspace(c))\r
 114                                 return NTT_String;\r
 115                         break;\r
 116 \r
 117                 case NTT_SingleQuotedString:\r
 118                         if ('\'' == c)\r
 119                                 return NTT_String;\r
 120                         break;\r
 121 \r
 122                 case NTT_DoubleQuotedString:\r
 123                         if ('"' == c)\r
 124                                 return NTT_String;\r
 125                         break;\r
 126 \r
 127                 case NTT_Comment:\r
 128                         if (']' == c)\r
 129                                 return GetToken(File, szToken, uBytes);\r
 130                         break;\r
 131 \r
 132                 default:\r
 133                         Quit("Tree::GetToken, invalid TT=%u", TT);\r
 134                         }\r
 135                 }\r
 136         }\r
 137 \r
 138 // NOTE: this hack must come after definition of Tree::GetToken.\r
 139 #if     TRACE\r
 140 #define GetToken        GetTokenVerbose\r
 141 #endif\r
 142 \r
 143 void Tree::FromFile(TextFile &File)\r
 144         {\r
 145 // Assume rooted.\r
 146 // If we discover that it is unrooted, will convert on the fly.\r
 147         CreateRooted();\r
 148 \r
 149         double dEdgeLength;\r
 150         bool bEdgeLength = GetGroupFromFile(File, 0, &dEdgeLength);\r
 151 \r
 152 // Next token should be either ';' for rooted tree or ',' for unrooted.\r
 153         char szToken[16];\r
 154         NEWICK_TOKEN_TYPE NTT = GetToken(File, szToken, sizeof(szToken));\r
 155 \r
 156 // If rooted, all done.\r
 157         if (NTT_Semicolon == NTT)\r
 158                 {\r
 159                 if (bEdgeLength)\r
 160                         Log(" *** Warning *** edge length on root group in Newick file %s\n",\r
 161                           File.GetFileName());\r
 162                 Validate();\r
 163                 return;\r
 164                 }\r
 165 \r
 166         if (NTT_Comma != NTT)\r
 167                 Quit("Tree::FromFile, expected ';' or ',', got '%s'", szToken);\r
 168 \r
 169         const unsigned uThirdNode = UnrootFromFile();\r
 170         bEdgeLength = GetGroupFromFile(File, uThirdNode, &dEdgeLength);\r
 171         if (bEdgeLength)\r
 172                 SetEdgeLength(0, uThirdNode, dEdgeLength);\r
 173         Validate();\r
 174         }\r
 175 \r
 176 // Return true if edge length for this group.\r
 177 bool Tree::GetGroupFromFile(TextFile &File, unsigned uNodeIndex,\r
 178   double *ptrdEdgeLength)\r
 179         {\r
 180         char szToken[1024];\r
 181         NEWICK_TOKEN_TYPE NTT = GetToken(File, szToken, sizeof(szToken));\r
 182 \r
 183 // Group is either leaf name or (left, right).\r
 184         if (NTT_String == NTT)\r
 185                 {\r
 186                 SetLeafName(uNodeIndex, szToken);\r
 187 #if     TRACE\r
 188                 Log("Group is leaf '%s'\n", szToken);\r
 189 #endif\r
 190                 }\r
 191         else if (NTT_Lparen == NTT)\r
 192                 {\r
 193                 const unsigned uLeft = AppendBranch(uNodeIndex);\r
 194                 const unsigned uRight = uLeft + 1;\r
 195 \r
 196         // Left sub-group...\r
 197 #if     TRACE\r
 198                 Log("Got '(', group is compound, expect left sub-group\n");\r
 199 #endif\r
 200                 double dEdgeLength;\r
 201                 bool bLeftLength = GetGroupFromFile(File, uLeft, &dEdgeLength);\r
 202 #if     TRACE\r
 203                 if (bLeftLength)\r
 204                         Log("Edge length for left sub-group: %.3g\n", dEdgeLength);\r
 205                 else\r
 206                         Log("No edge length for left sub-group\n");\r
 207 #endif\r
 208                 if (bLeftLength)\r
 209                         SetEdgeLength(uNodeIndex, uLeft, dEdgeLength);\r
 210 \r
 211         // ... then comma ...\r
 212 #if     TRACE\r
 213                 Log("Expect comma\n");\r
 214 #endif\r
 215                 NTT = GetToken(File, szToken, sizeof(szToken));\r
 216                 if (NTT_Comma != NTT)\r
 217                         Quit("Tree::GetGroupFromFile, expected ',', got '%s'", szToken);\r
 218 \r
 219         // ...then right sub-group...\r
 220 #if     TRACE\r
 221                 Log("Expect right sub-group\n");\r
 222 #endif\r
 223                 bool bRightLength = GetGroupFromFile(File, uRight, &dEdgeLength);\r
 224                 if (bRightLength)\r
 225                         SetEdgeLength(uNodeIndex, uRight, dEdgeLength);\r
 226 \r
 227 #if     TRACE\r
 228                 if (bRightLength)\r
 229                         Log("Edge length for right sub-group: %.3g\n", dEdgeLength);\r
 230                 else\r
 231                         Log("No edge length for right sub-group\n");\r
 232 #endif\r
 233 \r
 234         // ... then closing parenthesis.\r
 235 #if     TRACE\r
 236                 Log("Expect closing parenthesis (or comma if > 2-ary)\n");\r
 237 #endif\r
 238                 NTT = GetToken(File, szToken, sizeof(szToken));\r
 239                 if (NTT_Rparen == NTT)\r
 240                         ;\r
 241                 else if (NTT_Comma == NTT)\r
 242                         {\r
 243                         File.PushBack(',');\r
 244                         return false;\r
 245                         }\r
 246                 else\r
 247                         Quit("Tree::GetGroupFromFile, expected ')' or ',', got '%s'", szToken);\r
 248                 }\r
 249         else\r
 250                 Quit("Tree::GetGroupFromFile, expected '(' or leaf name, got '%s'",\r
 251                   szToken);\r
 252 \r
 253 // Group may optionally be followed by edge length.\r
 254         bool bEof = File.SkipWhiteX();\r
 255         if (bEof)\r
 256                 return false;\r
 257         char c;\r
 258         File.GetCharX(c);\r
 259 #if     TRACE\r
 260         Log("Character following group, could be colon, is '%c'\n", c);\r
 261 #endif\r
 262         if (':' == c)\r
 263                 {\r
 264                 NTT = GetToken(File, szToken, sizeof(szToken));\r
 265                 if (NTT_String != NTT)\r
 266                         Quit("Tree::GetGroupFromFile, expected edge length, got '%s'", szToken);\r
 267                 *ptrdEdgeLength = atof(szToken);\r
 268                 return true;\r
 269                 }\r
 270         File.PushBack(c);\r
 271         return false;\r
 272         }\r