[svn-r14892] RFC display compression ratio in h5dump

Here's the current behavior of h5dump regarding the printing of the dataset creation property list

For example

./h5dump -H -p -d filters

HDF5 "tfilters.h5" {
DATASET "deflate" {
   DATATYPE  H5T_STD_I32LE
   DATASPACE  SIMPLE { ( 20, 10 ) / ( 20, 10 ) }
   STORAGE_LAYOUT {
      CHUNKED ( 10, 5 )
      SIZE 385 
    }
   FILTERS {
      COMPRESSION DEFLATE { LEVEL 9 }
   }
   FILLVALUE {
      FILL_TIME H5D_FILL_TIME_IFSET
      VALUE  0
   }
   ALLOCATION_TIME {
      H5D_ALLOC_TIME_INCR
   }
}
}


The proposed behavior is to add this information after SIZE

SIZE 385 (51.9%COMPRESSION)

That percentage is obtained trough

Per = (b-a) / a

Where a = theoretical size obtained by multiplying datum size times number of elements
b =  size obtained with H5Dget_storage_size

The final print would look like

HDF5 "tfilters.h5" {
DATASET "deflate" {
   DATATYPE  H5T_STD_I32LE
   DATASPACE  SIMPLE { ( 20, 10 ) / ( 20, 10 ) }
   STORAGE_LAYOUT {
      CHUNKED ( 10, 5 )
      SIZE 385 (51.9%COMPRESSION)
    }
   FILTERS {
      COMPRESSION DEFLATE { LEVEL 9 }
   }
   FILLVALUE {
      FILL_TIME H5D_FILL_TIME_IFSET
      VALUE  0
   }
   ALLOCATION_TIME {
      H5D_ALLOC_TIME_INCR
   }
}
}

tested: windows, linux, solaris

6 files changed, 82 insertions, 11 deletions

diff --git a/tools/h5dump/h5dump.c b/tools/h5dump/h5dump.c
index 70d889f..dbac91d 100644
--- a/tools/h5dump/h5dump.c
+++ b/tools/h5dump/h5dump.c
@@ -2440,6 +2440,9 @@ static void dump_fill_value(hid_t dcpl,hid_t type_id, hid_t obj_id)
  *
  * Programmer:  pvn
  *
+ * Modifications: pvn, March 28, 2008
+ *   Add a COMPRESSION ration information for cases when filters are present
+ *
  *-------------------------------------------------------------------------
  */
 static void
@@ -2455,7 +2458,7 @@ dump_dcpl(hid_t dcpl_id,hid_t type_id, hid_t obj_id)
     unsigned         szip_pixels_per_block;
     hsize_t          chsize[64];     /* chunk size in elements */
     int              rank;           /* rank */
-    char                name[256];          /* external file name       */
+    char             name[256];      /* external file name       */
     off_t            offset;         /* offset of external file     */
     hsize_t          size;           /* size of external file   */
     H5D_fill_value_t fvstatus;
@@ -2467,11 +2470,12 @@ dump_dcpl(hid_t dcpl_id,hid_t type_id, hid_t obj_id)
     unsigned         j;
 
     storage_size=H5Dget_storage_size(obj_id);
+    nfilters = H5Pget_nfilters(dcpl_id);
     ioffset=H5Dget_offset(obj_id);
     next=H5Pget_external_count(dcpl_id);
     strcpy(f_name,"\0");
 
-    /*-------------------------------------------------------------------------
+   /*-------------------------------------------------------------------------
     * STORAGE_LAYOUT
     *-------------------------------------------------------------------------
     */
@@ -2490,7 +2494,72 @@ dump_dcpl(hid_t dcpl_id,hid_t type_id, hid_t obj_id)
             HDfprintf(stdout, ", %Hu", chsize[i]);
         printf(" %s\n", dump_header_format->dataspacedimend);
         indentation(indent + COL);
-        HDfprintf(stdout, "SIZE %Hu\n ", storage_size);
+
+
+        /* if there are filters, print a compression ratio */
+        if ( nfilters )
+        {
+
+            hid_t sid = H5Dget_space( obj_id );
+            hid_t tid = H5Dget_type( obj_id );
+            size_t datum_size = H5Tget_size( tid );
+            hsize_t dims[H5S_MAX_RANK];
+            int ndims = H5Sget_simple_extent_dims( sid, dims, NULL);  
+            hsize_t nelmts = 1;
+            hsize_t size;
+            double per = 0;
+            hssize_t a, b;
+            int ok = 0;
+
+            /* only print the compression ratio for these filters */
+            for ( i = 0; i < nfilters; i++) 
+            {
+                cd_nelmts = NELMTS(cd_values);
+                filtn = H5Pget_filter2(dcpl_id, (unsigned)i, &filt_flags, &cd_nelmts,
+                    cd_values, sizeof(f_name), f_name, NULL);
+                
+                switch (filtn) 
+                {
+                case H5Z_FILTER_DEFLATE:
+                case H5Z_FILTER_SZIP:
+                case H5Z_FILTER_NBIT:
+                case H5Z_FILTER_SCALEOFFSET:
+                    ok = 1;
+                    break;
+                }
+            }
+            
+            if (ndims && ok )
+            {
+                
+                for (i = 0; i < ndims; i++)
+                {
+                    nelmts *= dims[i];
+                }
+                size = nelmts * datum_size;
+
+                a = size; b = storage_size;
+                if (a!=0)
+                    per = (double) (b-a)/a;
+                
+                per = -per;
+                per *=100;
+
+                HDfprintf(stdout, "SIZE %Hu (%.1f%%COMPRESSION)\n ", storage_size, per);
+                
+            }
+            else
+                HDfprintf(stdout, "SIZE %Hu\n ", storage_size);
+
+
+            H5Sclose(sid);
+            H5Tclose(tid);
+            
+        }
+        else
+        {
+            HDfprintf(stdout, "SIZE %Hu\n ", storage_size);
+        }
 
         /*end indent */
         indent -= COL;
@@ -2563,14 +2632,16 @@ dump_dcpl(hid_t dcpl_id,hid_t type_id, hid_t obj_id)
     * FILTERS
     *-------------------------------------------------------------------------
     */
-    nfilters = H5Pget_nfilters(dcpl_id);
+    
 
     indentation(indent + COL);
     printf("%s %s\n", FILTERS, BEGIN);
     indent += COL;
 
-    if (nfilters) {
-        for (i=0; i<nfilters; i++) {
+    if (nfilters) 
+    {
+        for (i=0; i<nfilters; i++) 
+        {
             cd_nelmts = NELMTS(cd_values);
             filtn = H5Pget_filter2(dcpl_id, (unsigned)i, &filt_flags, &cd_nelmts,
                 cd_values, sizeof(f_name), f_name, NULL);
diff --git a/tools/testfiles/tallfilters.ddl b/tools/testfiles/tallfilters.ddl
index 282c085..43c27a9 100644
--- a/tools/testfiles/tallfilters.ddl
+++ b/tools/testfiles/tallfilters.ddl
@@ -7,7 +7,7 @@ DATASET "all" {
    DATASPACE  SIMPLE { ( 20, 10 ) / ( 20, 10 ) }
    STORAGE_LAYOUT {
       CHUNKED ( 10, 5 )
-      SIZE 458
+      SIZE 458 (42.8%COMPRESSION)
     }
    FILTERS {
       PREPROCESSING SHUFFLE
diff --git a/tools/testfiles/tdeflate.ddl b/tools/testfiles/tdeflate.ddl
index fc98432..cc5b49b 100644
--- a/tools/testfiles/tdeflate.ddl
+++ b/tools/testfiles/tdeflate.ddl
@@ -7,7 +7,7 @@ DATASET "deflate" {
    DATASPACE  SIMPLE { ( 20, 10 ) / ( 20, 10 ) }
    STORAGE_LAYOUT {
       CHUNKED ( 10, 5 )
-      SIZE 385
+      SIZE 385 (51.9%COMPRESSION)
     }
    FILTERS {
       COMPRESSION DEFLATE { LEVEL 9 }
diff --git a/tools/testfiles/tnbit.ddl b/tools/testfiles/tnbit.ddl
index a8bf549..2f932b5 100644
--- a/tools/testfiles/tnbit.ddl
+++ b/tools/testfiles/tnbit.ddl
@@ -7,7 +7,7 @@ DATASET "nbit" {
    DATASPACE  SIMPLE { ( 20, 10 ) / ( 20, 10 ) }
    STORAGE_LAYOUT {
       CHUNKED ( 10, 5 )
-      SIZE 76
+      SIZE 76 (90.5%COMPRESSION)
     }
    FILTERS {
       COMPRESSION NBIT
diff --git a/tools/testfiles/tscaleoffset.ddl b/tools/testfiles/tscaleoffset.ddl
index 6693f92..c9a46f4 100644
--- a/tools/testfiles/tscaleoffset.ddl
+++ b/tools/testfiles/tscaleoffset.ddl
@@ -7,7 +7,7 @@ DATASET "scaleoffset" {
    DATASPACE  SIMPLE { ( 20, 10 ) / ( 20, 10 ) }
    STORAGE_LAYOUT {
       CHUNKED ( 10, 5 )
-      SIZE 152
+      SIZE 152 (81.0%COMPRESSION)
     }
    FILTERS {
       COMPRESSION SCALEOFFSET { MIN BITS 4 }
diff --git a/tools/testfiles/tszip.ddl b/tools/testfiles/tszip.ddl
index 4c6af47..6411bc3 100644
--- a/tools/testfiles/tszip.ddl
+++ b/tools/testfiles/tszip.ddl
@@ -7,7 +7,7 @@ DATASET "szip" {
    DATASPACE  SIMPLE { ( 20, 10 ) / ( 20, 10 ) }
    STORAGE_LAYOUT {
       CHUNKED ( 10, 5 )
-      SIZE 372
+      SIZE 372 (53.5%COMPRESSION)
     }
    FILTERS {
       COMPRESSION SZIP {