Add algorithms by cat923 · Pull Request #243 · SparseLinearAlgebra/spla

cat923 · 2026-07-16T09:56:40Z

Implemented graph algorithms and test coverage for result comparison between Classic and SPLA versions.
Added project documentation with usage instructions and input format examples.
Added graph input readers for both vectors (.txt) and Matrix Market (.mtx) formats.

gsvgit · 2026-07-29T08:05:55Z

@@ -0,0 +1,16 @@
+# Install ruff to check your code style :


What is it for?

gsvgit · 2026-07-29T08:06:45Z

@@ -0,0 +1,21 @@
+from collections import deque
+
+INF = int(1e9)


Why you need distance and infinite in BFS?

gsvgit · 2026-07-29T08:08:37Z

+        depth += 1
+        count += front_size
+        v.assign(front, depth, op_assign=INT.SECOND, op_select=INT.NQZERO)
+        front = front.vxm(v, A, op_mult=INT.LAND, op_add=INT.LOR, op_select=INT.EQZERO)


How you prevent repeated visiting of vertices?

I copypasted this code from __init__.py. It says there is a mask to prevent repeating. I think it is op_select=INT.EQZERO

gsvgit · 2026-07-29T08:09:56Z

+    return graph, n
+
+
+def read_mtx_weighted(filename):


Is it helper for MTX reading? Can we use standard functions to do it?

Should i use mmread or is there a way to read MTX from SPLA ?

gsvgit · 2026-07-29T08:10:54Z

+from pyspla import FLOAT, INT, Matrix
+
+
+def read_mtx_int(filename):


Too many helpers with duplicated code. There are standard functions for mtx reading.

gsvgit · 2026-07-29T08:12:46Z

+    read_vectors_weighted,
+)
+from pr_classic import pagerank_classic
+from sssp_classic import INF as SSSP_INF


Yet another inf?

gsvgit · 2026-07-29T08:13:32Z

+from tc_spla import cohen
+
+
+def main():


Why you duplicate code?

gsvgit · 2026-07-29T08:14:42Z

+import math
+
+
+def pagerank_classic(adj_in, out_degree, n, alpha, eps):


What is the reason to implement it again? can you use existing reference implementation?

You told me to make a classic version for checking during one of our meetings .Could you please point out what exactly needs to be redone?

gsvgit · 2026-07-29T08:15:41Z

@@ -0,0 +1,14 @@
+INF = 1e9


It is not an infinity....

gsvgit · 2026-07-29T08:16:47Z

@@ -0,0 +1,41 @@
+from pyspla import FLOAT, Matrix, Scalar, Vector
+
+INF = float(1e9)


infinity can be represented in floats.

gsvgit · 2026-09-05T05:54:11Z

+
+def is_converged(old, new, eps=EPS):
+    diff = old.eadd(FLOAT.MINUS_POW2, new)
+    error2 = diff.reduce(FLOAT.PLUS)


It is strange that in SSSP we need to check error with epsilon.

I’ve fixed it

gsvgit · 2026-10-02T08:33:27Z

+    std::cout << std::endl;
    for (std::size_t k = 0; k < loader.get_n_values(); ++k) {
-        A->set_float(Ai[k], Aj[k], w);
+        A->set_float(Ai[k], Aj[k], Aw[k]);


Why these changes needed?

To use the actual edge weights loaded from the mtx file instead of the constant weight w = 1.

gsvgit · 2026-10-02T08:33:32Z


    for (std::size_t k = 0; k < loader.get_n_values(); ++k) {
-        if (Ai[k] > Aj[k]) {
+        if (Ai[k] < Aj[k]) {


Why these changes needed?

There is no difference in the results, but I like upper triangles more.

gsvgit · 2026-10-02T08:35:54Z


        [[nodiscard]] SPLA_API const std::vector<uint>& get_Ai() const;
        [[nodiscard]] SPLA_API const std::vector<uint>& get_Aj() const;
+        [[nodiscard]] SPLA_API const std::vector<float>& get_Aw() const;


Why these changes needed?

To use the actual edge weights loaded from the mtx file instead of the constant weight w = 1.

gsvgit · 2026-10-02T08:39:48Z

        std::size_t n_lines = 0;
        std::size_t n_sort  = 0;

+        std::string first_line;


What reasons of changes in this file?

Previously, all matrices were made symmetric by default, and weights were set to 1.0 for SSSP. Now, reverse edges are added only for symmetric matrices, and 1 weight is set specifically for pattern matrices.

cat923 added 6 commits June 12, 2026 20:02

Add algorithms

ffabf62

clean result files and fix SPLA algorithms

102eeff

Add README.md

0886e3a

Fix README.md

2ac847a

Add docstrings for spla algorithms and format files using ruff

b1f7bd6

Add ruff linter and formatter

5687bdd

gsvgit requested changes Jul 29, 2026

View reviewed changes

cat923 closed this Aug 4, 2026

cat923 reopened this Aug 17, 2026

Fix algorithms

e289c3f

cat923 requested a review from gsvgit August 18, 2026 15:02

cat923 added 3 commits August 18, 2026 21:30

Add scalar and rename pr to pr_spla

3c4519b

Add SSSP convergence

f72c245

remove unused idx variable

6ff8b32

gsvgit requested changes Sep 5, 2026

View reviewed changes

fix sssp_spla

8f461db

cat923 requested a review from gsvgit September 9, 2026 11:32

gsvgit approved these changes Sep 14, 2026

View reviewed changes

cat923 added 12 commits September 17, 2026 04:56

sssp_spla: add transpose

61996d0

delete using classic algorithms

bb71639

delete reading classic algorithms

5671372

fix graph.txt

dacac5b

fix tests

493fbcc

rename algorithms functions

99228d0

add set_fill_value to Vector and SSSP

bc3acfe

Add read weights

3bf140c

improved pr and sssp

971be16

Add sort

f4a59c3

use upper triangle

58e686c

Add edge weights and disable automatic symmetrization in MtxLoader

e8db2d0

cat923 added 3 commits October 1, 2026 01:12

Add fill_with for pr and set_fill_value for sssp

bcd01f4

Simplify Vector.set_fill_value to accept Scalar only

8e94ad7

Update tests

af7e938

cat923 requested a review from gsvgit September 30, 2026 22:55

Delete RUFF.md

db8c49b

gsvgit reviewed Oct 2, 2026

View reviewed changes

apply clang-format fixes

edfea2b

cat923 requested a review from gsvgit October 2, 2026 13:05

Delete stray endl in sssp example

63ed9a3

		from pyspla import FLOAT, INT, Matrix


		def read_mtx_int(filename):

		import math


		def pagerank_classic(adj_in, out_degree, n, alpha, eps):

		@@ -0,0 +1,41 @@
		from pyspla import FLOAT, Matrix, Scalar, Vector

		INF = float(1e9)

Conversation

cat923 commented Jul 16, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cat923 Sep 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

cat923 Sep 6, 2026 •

edited

Loading