KuangshiAi commited on
Commit
967a3ab
·
1 Parent(s): eee57cc

modify cases from Guoxi

Browse files
.DS_Store CHANGED
Binary files a/.DS_Store and b/.DS_Store differ
 
eval_cases/topology/topology_cases.yaml CHANGED
@@ -231,11 +231,11 @@
231
  assert:
232
  - type: rule_based
233
  eval_script: molecule/GS/molecule_eval.py
234
- eval_function:
235
  gs_file:
236
- - molecule/GS/molecule_gs.vtk
237
  rs_file:
238
- - molecule/results/{agent_mode}/molecule.vtk
239
 
240
 
241
  # 8. moons
@@ -253,9 +253,9 @@
253
  eval_script: moons/GS/moons_eval.py
254
  eval_function: evaluateMoonAscendingManifold
255
  gs_file:
256
- - moons/GS/moons_gs.vtk
257
  rs_file:
258
- - moons/results/{agent_mode}/moons.vtk
259
 
260
 
261
  # 9. dragon
@@ -276,19 +276,6 @@
276
  * 2 for 2-saddles
277
  * 3 for maxima
278
  - The point coordinates should be in world coordinates
279
-
280
- 5. Analyze the visualization and answer the following questions:
281
-
282
- Q1: How many saddle-maximum pairs are present in the dataset?
283
- (A) 2 (B) 4 (C) 6 (D) 10.
284
-
285
- Q2: How many minima are computed?
286
- (A) 2 (B) 5 (C) 8 (D) 10.
287
-
288
- Q3: Are there any saddle-saddle pairs in the persistence diagram?
289
- (Yes/No)
290
-
291
- Save the answers to the analysis questions in plain text as "dragon/results/{agent_mode}/answers.txt".
292
  assert:
293
  - type: rule_based
294
  eval_script: dragon/GS/dragon_eval.py
@@ -296,12 +283,4 @@
296
  gs_file:
297
  - dragon/GS/dragon_gs.vtk
298
  rs_file:
299
- - dragon/results/{agent_mode}/dragon.vtk
300
- - type: llm-rubric
301
- subtype: text
302
- value: |
303
- 1. Q1 correct answer: (C) 6
304
-
305
- 2. Q2 correct answer: (B) 5
306
-
307
- 3. Q3 correct answer: No
 
231
  assert:
232
  - type: rule_based
233
  eval_script: molecule/GS/molecule_eval.py
234
+ eval_function: evaluateMoleculeSegmentation
235
  gs_file:
236
+ - molecule/GS/molecule_gs.vti
237
  rs_file:
238
+ - molecule/results/{agent_mode}/molecule.vti
239
 
240
 
241
  # 8. moons
 
253
  eval_script: moons/GS/moons_eval.py
254
  eval_function: evaluateMoonAscendingManifold
255
  gs_file:
256
+ - moons/GS/moons_gs.vti
257
  rs_file:
258
+ - moons/results/{agent_mode}/moons.vti
259
 
260
 
261
  # 9. dragon
 
276
  * 2 for 2-saddles
277
  * 3 for maxima
278
  - The point coordinates should be in world coordinates
 
 
 
 
 
 
 
 
 
 
 
 
 
279
  assert:
280
  - type: rule_based
281
  eval_script: dragon/GS/dragon_eval.py
 
283
  gs_file:
284
  - dragon/GS/dragon_gs.vtk
285
  rs_file:
286
+ - dragon/results/{agent_mode}/dragon.vtk
 
 
 
 
 
 
 
 
topology/dragon/task_description.txt CHANGED
@@ -11,17 +11,4 @@
11
  * 1 for 1-saddles
12
  * 2 for 2-saddles
13
  * 3 for maxima
14
- - The point coordinates should be in world coordinates
15
-
16
- 5. Analyze the visualization and answer the following questions:
17
-
18
- Q1: How many saddle-maximum pairs are present in the dataset?
19
- (A) 2 (B) 4 (C) 6 (D) 10.
20
-
21
- Q2: How many minima are computed?
22
- (A) 2 (B) 5 (C) 8 (D) 10.
23
-
24
- Q3: Are there any saddle-saddle pairs in the persistence diagram?
25
- (Yes/No)
26
-
27
- Save the answers to the analysis questions in plain text as "dragon/results/{agent_mode}/answers.txt".
 
11
  * 1 for 1-saddles
12
  * 2 for 2-saddles
13
  * 3 for maxima
14
+ - The point coordinates should be in world coordinates
 
 
 
 
 
 
 
 
 
 
 
 
 
topology/molecule/GS/molecule_eval.py ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ import sys
3
+
4
+ # Add the topology directory to Python path
5
+ sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../..'))
6
+
7
+ from topologyScoring import partitionTopologicalDiceScore
8
+
9
+ def evaluateMoleculeSegmentation(gtFilename : str, reconFilename : str, verbose : bool = False) -> int:
10
+ """
11
+ Given two Morse-Smale segmentations of the same domain, return a similarity score from 0-10.
12
+ A score of 0 is considered bad and a score of 10 is considered good. The segmentations should be
13
+ represented by a point array called "Segmentation" that assigs a region identifier to each point in
14
+ the domain. The region identifiers between the ground truth and reconstructed files do not need to match.
15
+ Args:
16
+ gtFilename: The name of a file storing VTK image data (.vti) storing the ground truth MS segmentation.
17
+ each point's region ID should be stored in a point array called "Segmentation".
18
+ reconFilename: The name of a file storing VTK image data (.vti) storing the reconstructed MS segmentation.
19
+ verbose: Should error messages be printed if there are issues with the input files.
20
+ """
21
+
22
+ return partitionTopologicalDiceScore(gtFilename, "Segmentation", reconFilename, "Segmentation", verbose)
23
+
24
+ if __name__ == "__main__":
25
+
26
+ if len(sys.argv) != 3:
27
+ print(f"{os.path.basename(__file__)}: usage is 'python3 {os.path.basename(__file__)} gt_filename.vti recon_filename.vti")
28
+ exit(1)
29
+
30
+ score = evaluateMoleculeSegmentation(sys.argv[1], sys.argv[2], verbose=True)
31
+
32
+ print(f"This MS segmentation scored: {score}")