add a --sample option to 'pdal info', which is off by default for dat…

…a output #208
PDAL · Nov 18, 2013 · b1b2997 · b1b2997
1 parent ff54b7e
commit b1b2997
Show file tree

Hide file tree

Showing 3 changed files with 9 additions and 5 deletions.
diff --git a/include/pdal/kernel/Info.hpp b/include/pdal/kernel/Info.hpp
@@ -94,7 +94,8 @@ class PDAL_DLL Info : public Application
     std::string m_QueryPoint;
     double m_QueryDistance;
     boost::uint64_t m_numPointsToWrite;
-    std::string m_pipelineFile;    
+    std::string m_pipelineFile;
+    bool m_showSample;
 };
 
 }} // pdal::kernel

diff --git a/src/filters/Stats.cpp b/src/filters/Stats.cpp
@@ -497,6 +497,7 @@ void Stats::readBufferBeginImpl(PointBuffer& buffer)
         }
         else
         {
+            bool doSample = options.getValueOrDefault<bool>("do_sample", true);
             schema::index_by_index const& dims = schema.getDimensions().get<schema::index>();
             for (schema::index_by_index::const_iterator iter = dims.begin(); iter != dims.end(); ++iter)
             {
@@ -510,7 +511,7 @@ void Stats::readBufferBeginImpl(PointBuffer& buffer)
                     getStage().log()->get(logDEBUG2) << "Cumulating exact stats for dimension " << d->getName() << std::endl;
                     doExact = true;
                 }
-                stats::SummaryPtr c = boost::shared_ptr<stats::Summary>(new stats::Summary(bin_count, sample_size, stats_cache_size, seed, doExact));
+                stats::SummaryPtr c = boost::shared_ptr<stats::Summary>(new stats::Summary(bin_count, sample_size, stats_cache_size, seed, doExact, doSample));
 
                 std::pair<DimensionPtr, stats::SummaryPtr> p(d,c);
                 m_dimensions.push_back(d);

diff --git a/src/kernel/Info.cpp b/src/kernel/Info.cpp
@@ -51,6 +51,7 @@ Info::Info(int argc, const char* argv[])
     , m_useXML(false)
     , m_QueryDistance(0.0)
     , m_numPointsToWrite(0)
+    , m_showSample(false)
 {
     return;
 }
@@ -101,10 +102,11 @@ void Info::addSwitches()
         ("metadata,m", po::value<bool>(&m_showMetadata)->zero_tokens()->implicit_value(true), "dump the metadata")
         ("sdo_pc", po::value<bool>(&m_showSDOPCMetadata)->zero_tokens()->implicit_value(true), "dump the SDO_PC Oracle Metadata")
         ("stage,r", po::value<bool>(&m_showStage)->zero_tokens()->implicit_value(true), "dump the stage info")
+        ("pipeline-serialization", po::value<std::string>(&m_pipelineFile)->default_value(""), "")
         ("xml", po::value<bool>(&m_useXML)->zero_tokens()->implicit_value(true), "dump XML instead of JSON")
+        ("sample", po::value<bool>(&m_showSample)->zero_tokens()->implicit_value(true), "randomly sample dimension for stats")
         ("seed", po::value<boost::uint32_t>(&m_seed)->default_value(0), "Seed value for random sample")
         ("sample_size", po::value<boost::uint32_t>(&m_sample_size)->default_value(1000), "Sample size for random sample")
-        ("pipeline-serialization", po::value<std::string>(&m_pipelineFile)->default_value(""), "")
 
         ;
 
@@ -363,12 +365,12 @@ int Info::execute()
     m_options.add(cls);
     m_options.add(rn);
     m_options.add(nr);
+    Option do_sample("do_sample", m_showSample, "Dont do sampling");
+    m_options.add(do_sample);
     if (m_Dimensions.size())
     {
         Option dimensions("dimensions", m_Dimensions, "Use explicit list of dimensions");
         m_options.add(dimensions);
-        Option do_sample("do_sample", false, "Dont do sampling");
-        m_options.add(do_sample);
     }
 
     pdal::Options options = m_options + readerOptions;