diff --git a/README.md b/README.md
index 8f4f6d4..2c564be 100644
--- a/README.md
+++ b/README.md
@@ -33,3 +33,16 @@ conda develop install .
 #or with pip
 pip install --editable .
 ```
+
+## Cluster file specifications
+
+```
+[int32 frame_number][int32 n_clusters][clusters....]
+
+// Cluster data type
+typedef struct {
+  int16_t x;
+  int16_t y;
+  int32_t data[9];
+} Cluster ;
+```
\ No newline at end of file
diff --git a/src/cluster_reader.c b/src/cluster_reader.c
index f36e87a..1c2996b 100644
--- a/src/cluster_reader.c
+++ b/src/cluster_reader.c
@@ -1,87 +1,90 @@
 #include "cluster_reader.h"
 
-int read_clusters(FILE* fp, int64_t n_clusters, Cluster* buf, int *n_left){
-#ifdef CR_VERBOSE  
-  printf("Item size: %lu n_clusters: %lld, n_left: %d\n", sizeof(Cluster), n_clusters,*n_left);
+int read_clusters(FILE *fp, int64_t n_clusters, Cluster *buf, int *n_left) {
+#ifdef CR_VERBOSE
+    printf("Item size: %lu n_clusters: %lld, n_left: %d\n", sizeof(Cluster),
+           n_clusters, *n_left);
 #endif
-    int iframe=0, nph=*n_left;
-    size_t n_read=0, nph_read=0, nn=*n_left, nr=0;
-    //n_left=n_clusters;
-    
-    //read photons left from previous frame
+    int iframe = 0, nph = *n_left;
+    size_t n_read = 0, nph_read = 0, nn = *n_left, nr = 0;
+    // n_left=n_clusters;
+
+    // read photons left from previous frame
     if (nph) {
-	if (nph>n_clusters-nph_read)
-	  nn=n_clusters-nph_read;
-	else 
-	  nn=nph;
-	//printf("* %d %d %d %d\n",iframe,nph,nn,n_left);
-      	nr += fread((void*)(buf+nph_read), sizeof(Cluster), nn, fp);
-	n_read+=nr/sizeof(Cluster);
-	nph_read+=nn;
-	*n_left=nph-nn;
-    } 
-    if (nph_read<n_clusters) {
-    //keep on reading frames and photons until reaching n_clusters
-    while (fread(&iframe, sizeof(iframe), 1, fp)) {
-      if (fread(&nph, sizeof(nph), 1, fp) ) {
-	if (nph>n_clusters-nph_read)
-	  nn=n_clusters-nph_read;
-	else 
-	  nn=nph;
-	
-	//printf("%d %d %d %d\n",iframe,nph,nr,n_left);
-	nr += fread((void*)(buf+nph_read), sizeof(Cluster), nn, fp);
-	//printf("%d %d %d %d\n",iframe,nph,nr,n_left);
-	n_read+=nr;
-	nph_read+=nn;
-	*n_left=nph-nn;
-      }
-      if (nph_read>=n_clusters)
-	break;
+        if (nph > n_clusters - nph_read)
+            nn = n_clusters - nph_read;
+        else
+            nn = nph;
+        // printf("* %d %d %d %d\n",iframe,nph,nn,n_left);
+        nr += fread((void *)(buf + nph_read), sizeof(Cluster), nn, fp);
+        n_read += nr / sizeof(Cluster);
+        nph_read += nn;
+        *n_left = nph - nn;
     }
+    if (nph_read < n_clusters) {
+        // keep on reading frames and photons until reaching n_clusters
+        while (fread(&iframe, sizeof(iframe), 1, fp)) {
+            if (fread(&nph, sizeof(nph), 1, fp)) {
+                if (nph > n_clusters - nph_read)
+                    nn = n_clusters - nph_read;
+                else
+                    nn = nph;
+
+                // printf("%d %d %d %d\n",iframe,nph,nr,n_left);
+                nr += fread((void *)(buf + nph_read), sizeof(Cluster), nn, fp);
+                // printf("%d %d %d %d\n",iframe,nph,nr,n_left);
+                n_read += nr;
+                nph_read += nn;
+                *n_left = nph - nn;
+            }
+            if (nph_read >= n_clusters)
+                break;
+        }
     }
     // size_t n_read = fread(buf, sizeof(Cluster), n_clusters, fp);
 #ifdef CR_VERBOSE
-    printf("Read: %zu items %zu left %d\n", nph_read, n_read,*n_left);
+    printf("Read: %zu items %zu left %d\n", nph_read, n_read, *n_left);
 #endif
     return nph_read;
 }
-  
-int analyze_clusters(int64_t n_clusters, Cluster* cin, ClusterAnalysis *cout){
 
-  int32_t tot2[4], t2max;
-  char c;
-  int32_t val,tot3;
-  for (int ic=0; ic<n_clusters; ic++) {
-    tot3=0;
-    for (int i=0; i<4; i++)
-      tot2[i]=0;
-    //t2max=0;
-    for (int ix=0; ix<3; ix++) {
-      for (int iy=0; iy<3; iy++){
-	val=(cin+ic)->data[iy*3+ix];
-	tot3+=val;
-	if (ix<=1 && iy<=1) tot2[0]+=val;
-	if (ix>=1 && iy<=1) tot2[1]+=val;
-	if (ix<=1 && iy>=1) tot2[2]+=val;
-	if (ix>=1 && iy>=1) tot2[3]+=val;
-      }
-    } 
-    t2max=tot2[0];
-    c=cBottomLeft;
-    for (int i=1; i<4; i++) {
-      if (tot2[i]>t2max) {
-	t2max=tot2[i];
-	c=i;
-      }
+int analyze_clusters(int64_t n_clusters, Cluster *cin, ClusterAnalysis *cout) {
+
+    int32_t tot2[4], t2max;
+    char c;
+    int32_t val, tot3;
+    for (int ic = 0; ic < n_clusters; ic++) {
+        tot3 = 0;
+        for (int i = 0; i < 4; i++)
+            tot2[i] = 0;
+        // t2max=0;
+        for (int ix = 0; ix < 3; ix++) {
+            for (int iy = 0; iy < 3; iy++) {
+                val = (cin + ic)->data[iy * 3 + ix];
+                tot3 += val;
+                if (ix <= 1 && iy <= 1)
+                    tot2[0] += val;
+                if (ix >= 1 && iy <= 1)
+                    tot2[1] += val;
+                if (ix <= 1 && iy >= 1)
+                    tot2[2] += val;
+                if (ix >= 1 && iy >= 1)
+                    tot2[3] += val;
+            }
+        }
+        t2max = tot2[0];
+        c = cBottomLeft;
+        for (int i = 1; i < 4; i++) {
+            if (tot2[i] > t2max) {
+                t2max = tot2[i];
+                c = i;
+            }
+        }
+        (cout + ic)->c = c;
+        (cout + ic)->tot2 = t2max;
+        (cout + ic)->tot3 = tot3;
+        // printf("%d %d %d %d %d %d\n",ic,(cin+ic)->x, (cin+ic)->y,
+        // (cout+ic)->c, (cout+ic)->tot2, (cout+ic)->tot3);
     }
-    (cout+ic)->c=c;
-    (cout+ic)->tot2=t2max;
-    (cout+ic)->tot3=tot3;
-    //printf("%d %d %d %d %d %d\n",ic,(cin+ic)->x, (cin+ic)->y, (cout+ic)->c, (cout+ic)->tot2, (cout+ic)->tot3);
-    
-  }
-  return n_clusters;
-
-
+    return n_clusters;
 }
diff --git a/tests/test_ClusterReader.py b/tests/test_ClusterReader.py
index e0f9d48..325b69d 100644
--- a/tests/test_ClusterReader.py
+++ b/tests/test_ClusterReader.py
@@ -9,3 +9,20 @@ def test_references_on_read(data_path):
     clusters = r.read(10)
     assert sys.getrefcount(clusters) == 2 #Over counts by one due to call by reference
 
+def test_size_on_read(data_path):
+    fname= (data_path/'beam_En700eV_-40deg_300V_10us_d0_f0_100.clust').as_posix()
+    r = ClusterFileReader(fname)
+    
+    for i in range(10):
+        clusters = r.read(10)
+        assert clusters.size == 10
+
+def test_resize_on_read(data_path):
+    # File contains 481603 clusters, output should be resized to the correct size
+    fname= (data_path/'beam_En700eV_-40deg_300V_10us_d0_f0_100.clust').as_posix()
+    r = ClusterFileReader(fname)
+    
+    max_clusters = 10000000 #400MB initial allocation
+    clusters = r.read(max_clusters)
+    assert clusters.size == 481603
+    assert sys.getrefcount(clusters) == 2
\ No newline at end of file