From 0c3ebbfc485c4f09a7ba224270926ffe70f4ff2b Mon Sep 17 00:00:00 2001
From: Mark Keller <7525285+keller-mark@users.noreply.github.com>
Date: Thu, 11 Jun 2020 21:49:22 -0400
Subject: [PATCH] bbi.fetch with summary="sum" (#12)

* Sum aggregation mode

* Missed a line

* Test

* Use higher zoom level in test

* Actually do sum in test

* Improve test

* Whitespace
---
 bbi/__init__.py   | 2 +-
 bbi/cbbi.pxd      | 1 +
 bbi/cbbi.pyx      | 4 ++++
 include/bbiFile.h | 1 +
 src/bbiRead.c     | 7 +++++++
 tests/test_bbi.py | 4 ++++
 6 files changed, 18 insertions(+), 1 deletion(-)

diff --git a/bbi/__init__.py b/bbi/__init__.py
index 69dece9..2e9b1a9 100644
--- a/bbi/__init__.py
+++ b/bbi/__init__.py
@@ -15,4 +15,4 @@
 __all__ = ['is_bbi', 'is_bigwig', 'is_bigbed', 'info', 'zooms', 'chromsizes',
            'fetch', 'stackup', 'fetch_intervals']
 
-__version__ = '0.2.2'
+__version__ = '0.2.3'
diff --git a/bbi/cbbi.pxd b/bbi/cbbi.pxd
index f5fcf02..94122f3 100644
--- a/bbi/cbbi.pxd
+++ b/bbi/cbbi.pxd
@@ -122,6 +122,7 @@ cdef extern from "bbiFile.h":
         bbiSumMin = 2
         bbiSumCoverage = 3
         bbiSumStandardDeviation = 4
+        bbiSumSum = 5
 
     cdef struct bbiSummary:
         bbiSummary *next
diff --git a/bbi/cbbi.pyx b/bbi/cbbi.pyx
index aeaddea..7856618 100644
--- a/bbi/cbbi.pyx
+++ b/bbi/cbbi.pyx
@@ -78,6 +78,7 @@ cpdef dict BBI_SUMMARY_TYPES = {
     'min': bbiSumMin,
     'cov': bbiSumCoverage,
     'std': bbiSumStandardDeviation,
+    'sum': bbiSumSum,
 }
 
 
@@ -254,6 +255,7 @@ def info(str inFile):
             'std': sqrt(var_from_sums(summ.sumData,
                                       summ.sumSquares,
                                       summ.validCount)),
+            'sum': summ.sumData,
         }
     }
     bbiFileClose(&bbi)
@@ -731,6 +733,8 @@ cdef inline void array_query_summarized(
                         val = sqrt(var_from_sums(el.sumData,
                                                  el.sumSquares,
                                                  el.validCount))
+                    elif summaryType == bbiSumSum:
+                        val = el.sumData
                     else:
                         raise RuntimeError
                     out[i] = val
diff --git a/include/bbiFile.h b/include/bbiFile.h
index 0746afa..0dd3b66 100644
--- a/include/bbiFile.h
+++ b/include/bbiFile.h
@@ -200,6 +200,7 @@ enum bbiSummaryType
     bbiSumMin = 2,	/* Minimum value */
     bbiSumCoverage = 3,  /* Bases in region containing actual data. */
     bbiSumStandardDeviation = 4, /* Standard deviation in window. */
+    bbiSumSum = 5, /* Sum value */
     };
 
 enum bbiSummaryType bbiSummaryTypeFromString(char *string);
diff --git a/src/bbiRead.c b/src/bbiRead.c
index caf2d30..9f1d153 100644
--- a/src/bbiRead.c
+++ b/src/bbiRead.c
@@ -280,6 +280,8 @@ else if (sameWord(string, "coverage") || sameWord(string, "dataCoverage"))
     return bbiSumCoverage;
 else if (sameWord(string, "std"))
     return bbiSumStandardDeviation;
+else if (sameWord(string, "sum"))
+    return bbiSumSum;
 else
     {
     errAbort("Unknown bbiSummaryType %s", string);
@@ -302,6 +304,8 @@ switch (type)
         return "coverage";
     case bbiSumStandardDeviation:
         return "std";
+    case bbiSumSum:
+        return "sum";
     default:
 	errAbort("Unknown bbiSummaryType %d", (int)type);
 	return NULL;
@@ -691,6 +695,9 @@ if (ret)
 		case bbiSumStandardDeviation:
 		    val = calcStdFromSums(el->sumData, el->sumSquares, el->validCount);
 		    break;
+		case bbiSumSum:
+		    val = el->sumData;
+		    break;
 		default:
 		    internalErr();
 		    val = 0.0;
diff --git a/tests/test_bbi.py b/tests/test_bbi.py
index bad7424..8c7f24f 100644
--- a/tests/test_bbi.py
+++ b/tests/test_bbi.py
@@ -120,6 +120,10 @@ def test_fetch_summary_stats(path):
         path, 'chr21', 20000000, 20001000, bins=10, summary='max'
     ).max()
     assert np.isclose(vmax, np.max(values))
+    vsum = bbi.fetch(path, 'chr21', 20000000, 20001000, bins=100, summary='sum')
+    values_sum_every_ten = np.reshape(values, (-1, 10)).sum(axis=-1)
+    assert len(vsum) == len(values_sum_every_ten)
+    assert np.allclose(vsum, values_sum_every_ten)
 
     with pytest.raises(ValueError):
         bbi.fetch(path, 'chr21', 20000000, 20001000, bins=10, summary='foo')