Merge branch 'mb/features' into develop

1 files changed, 56 insertions, 0 deletions

diff --git a/bsfs/triple_store/sparql/distance.py b/bsfs/triple_store/sparql/distance.py
new file mode 100644
index 0000000..2f5387a
--- /dev/null
+++ b/bsfs/triple_store/sparql/distance.py
@@ -0,0 +1,56 @@
+"""
+
+Part of the BlackStar filesystem (bsfs) module.
+A copy of the license is provided with the project.
+Author: Matthias Baumgartner, 2022
+"""
+# standard imports
+import typing
+
+# external imports
+import numpy as np
+
+# bsfs imports
+from bsfs.namespace import ns
+
+# constants
+EPS = 1e-9
+
+# exports
+__all__: typing.Sequence[str] = (
+    'DISTANCE_FU',
+    )
+
+
+## code ##
+
+def euclid(fst, snd) -> float:
+    """Euclidean distance (l2 norm)."""
+    fst = np.array(fst)
+    snd = np.array(snd)
+    return float(np.linalg.norm(fst - snd))
+
+def cosine(fst, snd) -> float:
+    """Cosine distance."""
+    fst = np.array(fst)
+    snd = np.array(snd)
+    if (fst == snd).all():
+        return 0.0
+    nrm0 = np.linalg.norm(fst)
+    nrm1 = np.linalg.norm(snd)
+    return float(1.0 - np.dot(fst, snd) / (nrm0 * nrm1 + EPS))
+
+def manhatten(fst, snd) -> float:
+    """Manhatten (cityblock) distance (l1 norm)."""
+    fst = np.array(fst)
+    snd = np.array(snd)
+    return float(np.abs(fst - snd).sum())
+
+# Known distance functions.
+DISTANCE_FU = {
+    ns.bsfs.euclidean: euclid,
+    ns.bsfs.cosine: cosine,
+    ns.bsfs.manhatten: manhatten,
+}
+
+## EOF ##