Add GT URLs for wikipedia runbooks (#316)

* added gt for wiki runbooks * undo local change * fixed url for wiki-1M runbook
harsha-simhadri · Nov 1, 2024 · 541e3d1 · 541e3d1
1 parent 3c35f54
commit 541e3d1
Show file tree

Hide file tree

Showing 8 changed files with 54 additions and 18 deletions.
diff --git a/data_export.py b/data_export.py
@@ -96,18 +96,22 @@ def cleaned_run_metric(run_metrics):
             dataset = DATASETS[dataset_name]()
             runbook_paths = [None]
             if track == 'streaming':
-                runbook_paths = ['neurips23/runbooks/streaming/simple_runbook.yaml',
-                                    'neurips23/runbooks/streaming/simple_replace_runbook.yaml',
-                                    'neurips23/runbooks/streaming/random_replace_runbook.yaml',
-                                    'neurips23/runbooks/streaming/clustered_replace_runbook.yaml',
-                                    'neurips23/runbooks/streaming/clustered_runbook.yaml',
-                                    'neurips23/runbooks/streaming/clustered_runbook.yaml',
-                                    'neurips23/runbooks/streaming/delete_runbook.yaml',
-                                    'neurips23/runbooks/streaming/final_runbook.yaml',
-                                    'neurips23/runbooks/streaming/msturing-10M_slidingwindow_runbook.yaml',
-                                    'neurips23/runbooks/streaming/wikipedia-35M_expirationtime_runbook.yaml',
-                                    'neurips23/runbooks/streaming/wikipedia-35M_expiration_time_replace_runbook.yaml',
-                                    'neurips23/runbooks/streaming/msmarco-100M_expirationtime_runbook.yaml']
+                runbook_paths = ['neurips23/runbooks/simple_runbook.yaml',
+                                    'neurips23/runbooks/simple_replace_runbook.yaml',
+                                    'neurips23/runbooks/random_replace_runbook.yaml',
+                                    'neurips23/runbooks/clustered_replace_runbook.yaml',
+                                    'neurips23/runbooks/clustered_runbook.yaml',
+                                    'neurips23/runbooks/clustered_runbook.yaml',
+                                    'neurips23/runbooks/delete_runbook.yaml',
+                                    'neurips23/runbooks/final_runbook.yaml',
+                                    'neurips23/runbooks/msturing-10M_slidingwindow_runbook.yaml',
+                                    'neurips23/runbooks/wikipedia-35M_expirationtime_runbook.yaml',
+                                    'neurips23/runbooks/wikipedia-1M_expiration_time_runbook.yaml',
+                                    'neurips23/runbooks/wikipedia-35M_expiration_time_replace_only_runbook.yaml',
+                                    'neurips23/runbooks/wikipedia-1M_expiration_time_replace_only_runbook.yaml',
+                                    'neurips23/runbooks/wikipedia-35M_expiration_time_replace_delete_runbook.yaml',
+                                    'neurips23/runbooks/wikipedia-1M_expiration_time_replace_delete_runbook.yaml',
+                                    'neurips23/runbooks/msmarco-100M_expirationtime_runbook.yaml']
             for runbook_path in runbook_paths:
                 print("Looking for runbook ", runbook_path)
                 results = load_all_results(dataset_name, neurips23track=track, runbook_path=runbook_path)

diff --git a/neurips23/runbooks/gen_expiration_time_runbook.py b/neurips23/runbooks/gen_expiration_time_runbook.py
@@ -124,7 +124,7 @@ def gen_exp_time_runbook(dataset_name, dataset_size, max_t, runbook_filename, ra
 dataset_name = 'wikipedia-1M'
 dataset_size = 1000000
 max_t = 100
-gt_url = "https://comp21storage.z5.web.core.windows.net/wiki-cohere-35M/wikipedia-1M_expiration_time_runbook.yaml/"
+gt_url = "https://comp21storage.z5.web.core.windows.net/wiki-cohere-35M/wikipedia-1M_expiration_time_runbook.yaml"
 gen_exp_time_runbook(dataset_name, dataset_size, max_t, dataset_file, ratios, timesteps, seed, False, gt_url)
 
 ratios = (0, 4, 18)
@@ -134,7 +134,7 @@ def gen_exp_time_runbook(dataset_name, dataset_size, max_t, runbook_filename, ra
 dataset_name = 'wikipedia-35M'
 dataset_size = 8000000 #only use a prefix of the dataset
 max_t = 80
-gt_url = None
+gt_url = "https://comp21storage.z5.web.core.windows.net/wiki-cohere-35M/wikipedia-35M_expiration_time_replace_only_runbook.yaml"
 gen_exp_time_runbook(dataset_name, dataset_size, max_t, dataset_file, ratios, timesteps, seed, True, gt_url, False)
 
 ratios = (0, 4, 18)
@@ -144,7 +144,7 @@ def gen_exp_time_runbook(dataset_name, dataset_size, max_t, runbook_filename, ra
 dataset_name = 'wikipedia-1M'
 dataset_size = 1000000
 max_t = 100
-gt_url = None
+gt_url = "https://comp21storage.z5.web.core.windows.net/wiki-cohere-35M/wikipedia-1M_expiration_time_replace_only_runbook.yaml"
 gen_exp_time_runbook(dataset_name, dataset_size, max_t, dataset_file, ratios, timesteps, seed, True, gt_url, False)
 
 ratios = (3, 8, 18)
@@ -154,7 +154,8 @@ def gen_exp_time_runbook(dataset_name, dataset_size, max_t, runbook_filename, ra
 dataset_name = 'wikipedia-35M'
 dataset_size = 35000000
 max_t = 350
-gen_exp_time_runbook(dataset_name, dataset_size, max_t, dataset_file, ratios, timesteps, seed, True, None)
+gt_url = "https://comp21storage.z5.web.core.windows.net/wiki-cohere-35M/wikipedia-35M_expiration_time_replace_delete_runbook.yaml"
+gen_exp_time_runbook(dataset_name, dataset_size, max_t, dataset_file, ratios, timesteps, seed, True, gt_url)
 
 ratios = (1, 8, 18)
 timesteps = (0, 100, 20)
@@ -163,7 +164,8 @@ def gen_exp_time_runbook(dataset_name, dataset_size, max_t, runbook_filename, ra
 dataset_name = 'wikipedia-1M'
 dataset_size = 1000000
 max_t = 100
-gen_exp_time_runbook(dataset_name, dataset_size, max_t, dataset_file, ratios, timesteps, seed, True, None)
+gt_url = "https://comp21storage.z5.web.core.windows.net/wiki-cohere-35M/wikipedia-1M_expiration_time_replace_delete_runbook.yaml"
+gen_exp_time_runbook(dataset_name, dataset_size, max_t, dataset_file, ratios, timesteps, seed, True, gt_url)
 
 ratios = (0, 6, 25)
 timesteps = (0, 200, 50)

diff --git a/neurips23/runbooks/wikipedia-1M_expiration_time_replace_delete_runbook.yaml b/neurips23/runbooks/wikipedia-1M_expiration_time_replace_delete_runbook.yaml
@@ -1208,3 +1208,4 @@ wikipedia-1M:
   316:
     operation: search
   max_pts: 293233
+  gt_url: "https://comp21storage.z5.web.core.windows.net/wiki-cohere-35M/wikipedia-1M_expiration_time_replace_delete_runbook.yaml"
diff --git a/neurips23/runbooks/wikipedia-1M_expiration_time_replace_only_runbook.yaml b/neurips23/runbooks/wikipedia-1M_expiration_time_replace_only_runbook.yaml
@@ -1068,3 +1068,4 @@ wikipedia-1M:
   278:
     operation: search
   max_pts: 698369
+  gt_url: "https://comp21storage.z5.web.core.windows.net/wiki-cohere-35M/wikipedia-1M_expiration_time_replace_only_runbook.yaml"
diff --git a/neurips23/runbooks/wikipedia-1M_expiration_time_runbook.yaml b/neurips23/runbooks/wikipedia-1M_expiration_time_runbook.yaml
@@ -840,4 +840,4 @@ wikipedia-1M:
   260:
     operation: search
   max_pts: 410000
-  gt_url: https://comp21storage.z5.web.core.windows.net/wiki-cohere-35M/wikipedia-1M_expiration_time_runbook.yaml/
+  gt_url: https://comp21storage.z5.web.core.windows.net/wiki-cohere-35M/wikipedia-1M_expiration_time_runbook.yaml
diff --git a/neurips23/runbooks/wikipedia-35M_expiration_time_replace_delete_runbook.yaml b/neurips23/runbooks/wikipedia-35M_expiration_time_replace_delete_runbook.yaml
@@ -4436,3 +4436,4 @@ wikipedia-35M:
   1150:
     operation: search
   max_pts: 6682767
+  gt_url: "https://comp21storage.z5.web.core.windows.net/wiki-cohere-35M/wikipedia-35M_expiration_time_replace_delete_runbook.yaml"
diff --git a/neurips23/runbooks/wikipedia-35M_expiration_time_replace_only_runbook.yaml b/neurips23/runbooks/wikipedia-35M_expiration_time_replace_only_runbook.yaml
@@ -852,3 +852,4 @@ wikipedia-35M:
   222:
     operation: search
   max_pts: 5548955
+  gt_url: "https://comp21storage.z5.web.core.windows.net/wiki-cohere-35M/wikipedia-35M_expiration_time_replace_only_runbook.yaml"
diff --git a/neurips23/streaming/diskann/config.yaml b/neurips23/streaming/diskann/config.yaml
@@ -48,6 +48,32 @@ msturing-1M:
           query-args: |
             [{"Ls":300, "T":16},
              {"Ls":100, "T":16}]
+wikipedia-1M:
+    diskann: 
+      docker-tag: neurips23-streaming-diskann
+      module: neurips23.streaming.diskann.diskann-str
+      constructor: diskann
+      base-args: ["@metric"]
+      run-groups:
+        base:
+          args: |
+            [{"R":32, "L":100, "insert_threads":32, "consolidate_threads":32}]
+          query-args: |
+            [
+             {"Ls":100, "T":32}]
+wikipedia-35M:
+    diskann: 
+      docker-tag: neurips23-streaming-diskann
+      module: neurips23.streaming.diskann.diskann-str
+      constructor: diskann
+      base-args: ["@metric"]
+      run-groups:
+        base:
+          args: |
+            [{"R":32, "L":100, "insert_threads":32, "consolidate_threads":32}]
+          query-args: |
+            [
+             {"Ls":100, "T":32}]
 msspacev-10M:
     diskann: 
       docker-tag: neurips23-streaming-diskann