12d0f12bd304787c52cab0780e367d36b020f84e
kent
  Tue Feb 26 12:11:18 2013 -0800
Adding name index to bigBed files.  The write side I _think_ is working.  Still developing read side.
diff --git src/inc/bbiFile.h src/inc/bbiFile.h
index 949624d..2d31caa 100644
--- src/inc/bbiFile.h
+++ src/inc/bbiFile.h
@@ -8,61 +8,62 @@
 #include "localmem.h"
 
 /* bigWig/bigBed file structure:
  *     fixedWidthHeader
  *         magic# 		4 bytes
  *         version              2 bytes
  *	   zoomLevels		2 bytes
  *         chromosomeTreeOffset	8 bytes
  *         fullDataOffset	8 bytes
  *	   fullIndexOffset	8 bytes
  *         fieldCount           2 bytes (for bigWig 0)
  *         definedFieldCount    2 bytes (for bigWig 0)
  *         autoSqlOffset        8 bytes (for bigWig 0) (0 if no autoSql information)
  *         totalSummaryOffset   8 bytes (0 in earlier versions of file lacking totalSummary)
  *         uncompressBufSize    4 bytes (Size of uncompression buffer.  0 if uncompressed.)
- *         reserved             8 bytes (0 for now)
+ *         nameIndexOffset      8 bytes (Offset to name index, 0 if no such index)
  *     zoomHeaders		there are zoomLevels number of these
  *         reductionLevel	4 bytes
  *	   reserved		4 bytes
  *	   dataOffset		8 bytes
  *         indexOffset          8 bytes
  *     autoSql string (zero terminated - only present if autoSqlOffset non-zero)
  *     totalSummary - summary of all data in file - only present if totalSummaryOffset non-zero
  *         basesCovered        8 bytes
  *         minVal              8 bytes float (for bigBed minimum depth of coverage)
  *         maxVal              8 bytes float (for bigBed maximum depth of coverage)
  *         sumData             8 bytes float (for bigBed sum of coverage)
  *         sumSquared          8 bytes float (for bigBed sum of coverage squared)
  *     chromosome b+ tree       bPlusTree index
  *     full data
  *         sectionCount		8 bytes (item count for bigBeds)
  *         section data		section count sections, of three types (bed data for bigBeds)
  *     full index               cirTree index
  *     zoom info             one of these for each zoom level
  *         zoom data
  *             zoomCount	4 bytes
  *             zoom data	there are zoomCount of these items
  *                 chromId	4 bytes
  *	           chromStart	4 bytes
  *                 chromEnd     4 bytes
  *                 validCount	4 bytes
  *                 minVal       4 bytes float 
  *                 maxVal       4 bytes float
  *                 sumData      4 bytes float
  *                 sumSquares   4 bytes float
  *         zoom index        	cirTree index
+ *     name index [optional]    bPlusTree index
  *     magic# 		4 bytes - same as magic number at start of header
  */
 
 #ifndef CIRTREE_H
 #include "cirTree.h"
 #endif
 
 #define bbiCurrentVersion 4
 /* Version history (of file format, not utilities - corresponds to version field in header)
  *    1 - Initial release
  *    1 - Unfortunately when attempting a transparent change to encoders, made the sectionCount 
  *        field inconsistent, sometimes not present, sometimes 32 bits.  Since offset positions
  *        in index were still accurate this did not break most applications, but it did show
  *        up in the summary section of the Table Browser.
  *    2 - Made sectionCount consistently 64 bits. Also fixed missing zoomCount in first level of
@@ -94,32 +95,34 @@
     char *fileName;		/* Name of file - for better error reporting. */
     struct udcFile *udc;	/* Open UDC file handle. */
     bits32 typeSig;		/* bigBedSig or bigWigSig for now. */
     boolean isSwapped;		/* If TRUE need to byte swap everything. */
     struct bptFile *chromBpt;	/* Index of chromosomes. */
     bits16 version;		/* Version number - initially 1. */
     bits16 zoomLevels;		/* Number of zoom levels. */
     bits64 chromTreeOffset;	/* Offset to chromosome index. */
     bits64 unzoomedDataOffset;	/* Start of unzoomed data. */
     bits64 unzoomedIndexOffset;	/* Start of unzoomed index. */
     bits16 fieldCount;		/* Number of columns in bed version. */
     bits16 definedFieldCount;   /* Number of columns using bed standard definitions. */
     bits64 asOffset;		/* Offset to embedded null-terminated AutoSQL file. */
     bits64 totalSummaryOffset;	/* Offset to total summary information if any.  (On older files have to calculate) */
     bits32 uncompressBufSize;	/* Size of uncompression buffer, 0 if uncompressed */
+    bits64 nameIndexOffset;	/* Start of name index or zero if none. */
     struct cirTreeFile *unzoomedCir;	/* Unzoomed data index in memory - may be NULL. */
     struct bbiZoomLevel *levelList;	/* List of zoom levels. */
+    struct bptFile *nameBpt;	/* Index of names, may be NULL */
     };
 
 
 struct bbiFile *bbiFileOpen(char *fileName, bits32 sig, char *typeName);
 /* Open up big wig or big bed file. */
 
 void bbiFileClose(struct bbiFile **pBwf);
 /* Close down a big wig/big bed file. */
 
 struct fileOffsetSize *bbiOverlappingBlocks(struct bbiFile *bbi, struct cirTreeFile *ctf,
 	char *chrom, bits32 start, bits32 end, bits32 *retChromId);
 /* Fetch list of file blocks that contain items overlapping chromosome range. */
  
 struct bbiChromIdSize
 /* We store an id/size pair in chromBpt bPlusTree */
@@ -315,33 +318,33 @@
 void bbiWriteFloat(FILE *f, float val);
 /* Write out floating point val to file.  Mostly to convert from double... */
 
 struct hash *bbiChromSizesFromFile(char *fileName);
 /* Read two column file into hash keyed by chrom. */
 
 bits64 bbiTotalSummarySize(struct bbiSummary *list);
 /* Return size on disk of all summaries. */
 
 void bbiChromUsageFree(struct bbiChromUsage **pUsage);
 /* free a single bbiChromUsage structure */
 
 void bbiChromUsageFreeList(struct bbiChromUsage **pList);
 /* free a list of bbiChromUsage structures */
 
-struct bbiChromUsage *bbiChromUsageFromBedFile(struct lineFile *lf, 
-	struct hash *chromSizesHash, int *retMinDiff, double *retAveSize, bits64 *retBedCount);
-/* Go through bed file and collect chromosomes and statistics. Free with bbiChromUsageFreeList */
+struct bbiChromUsage *bbiChromUsageFromBedFile(struct lineFile *lf, struct hash *chromSizesHash, 
+	int *retMinDiff, double *retAveSize, bits64 *retBedCount, int *retMaxNameSize);
+/* Go through bed file and collect chromosomes and statistics. */
 
 int bbiCountSectionsNeeded(struct bbiChromUsage *usageList, int itemsPerSlot);
 /* Count up number of sections needed for data. */
 
 void bbiAddToSummary(bits32 chromId, bits32 chromSize, bits32 start, bits32 end, 
 	bits32 validCount, double minVal, double maxVal, double sumData, double sumSquares,  
 	int reduction, struct bbiSummary **pOutList);
 /* Add data range to summary - putting it onto top of list if possible, otherwise
  * expanding list. */
 
 void bbiAddRangeToSummary(bits32 chromId, bits32 chromSize, bits32 start, bits32 end, 
 	double val, int reduction, struct bbiSummary **pOutList);
 /* Add chromosome range to summary - putting it onto top of list if possible, otherwise
  * expanding list. */