Apply custom patch only once by comparing the last patch id

Signed-off-by: Heemin Kim <[email protected]>

.github/workflows/CI.yml

@@ -37,8 +37,6 @@ jobs:
     steps:
       - name: Checkout k-NN
         uses: actions/checkout@v1
-        with:
-          submodules: true
 
       # Setup git user so that patches for native libraries can be applied and committed
       - name: Setup git user

.github/workflows/test_security.yml

@@ -37,8 +37,6 @@ jobs:
     steps:
       - name: Checkout k-NN
         uses: actions/checkout@v1
-        with:
-          submodules: true
       # Setup git user so that patches for native libraries can be applied and committed
       - name: Setup git user
         run: |

CHANGELOG.md

@@ -23,6 +23,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 * Release memory properly for an array type [#1820](https://github.com/opensearch-project/k-NN/pull/1820)
 * FIX Same Suffix Cause Recall Drop to zero [#1802](https://github.com/opensearch-project/k-NN/pull/1802)
 ### Infrastructure
+* Apply custom patch only once by comparing the last patch id  [#1833](https://github.com/opensearch-project/k-NN/pull/1833)
 ### Documentation
 * Update dev guide to fix clang linking issue on arm [#1746](https://github.com/opensearch-project/k-NN/pull/1746)
 ### Maintenance

jni/cmake/init-faiss.cmake

@@ -12,11 +12,25 @@ if (NOT EXISTS ${FAISS_REPO_DIR})
     execute_process(COMMAND git submodule update --init -- external/faiss WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR})
 endif ()
 
-# Check if patch exist, this is to skip git apply during CI build. See CI.yml with ubuntu.
-find_path(PATCH_FILE NAMES 0001-Custom-patch-to-support-multi-vector.patch 0002-Enable-precomp-table-to-be-shared-ivfpq.patch 0003-Custom-patch-to-support-range-search-params.patch 0004-Custom-patch-to-support-binary-vector.patch PATHS ${CMAKE_CURRENT_SOURCE_DIR}/patches/faiss NO_DEFAULT_PATH)
+# Get the last patch file name
+file(GLOB PATCH_FILE_LIST "${CMAKE_CURRENT_SOURCE_DIR}/patches/faiss/*")
+list(SORT PATCH_FILE_LIST ORDER DESCENDING)
+list(GET PATCH_FILE_LIST 0 LAST_PATCH_FILE)
+message(STATUS "Last patch file: ${PATCH_FILE_LIST}")
 
-# If it exists, apply patches
-if (EXISTS ${PATCH_FILE})
+# Get patch id of the last patch
+execute_process(COMMAND sh -c "cat ${LAST_PATCH_FILE} | git patch-id --stable" OUTPUT_VARIABLE PATCH_ID_OUTPUT)
+string(REPLACE " " ";" PATCH_ID_LIST ${PATCH_ID_OUTPUT})
+list(GET PATCH_ID_LIST 0 PATCH_ID)
+message(STATUS "Extracted patch ID: ${PATCH_ID}")
+
+# Get patch id of the last commit
+execute_process(COMMAND sh -c "git --no-pager show HEAD | git patch-id --stable" OUTPUT_VARIABLE PATCH_ID_OUTPUT_FROM_COMMIT WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/external/faiss)
+string(REPLACE " " ";" PATCH_ID_LIST_FROM_COMMIT ${PATCH_ID_OUTPUT_FROM_COMMIT})
+list(GET PATCH_ID_LIST_FROM_COMMIT 0 PATCH_ID_FROM_COMMIT)
+message(STATUS "Extracted patch ID from commit: ${PATCH_ID_FROM_COMMIT}")
+
+if (NOT ${PATCH_ID} STREQUAL ${PATCH_ID_FROM_COMMIT})
     message(STATUS "Applying custom patches.")
     execute_process(COMMAND git ${GIT_PATCH_COMMAND} --3way --ignore-space-change --ignore-whitespace ${CMAKE_CURRENT_SOURCE_DIR}/patches/faiss/0001-Custom-patch-to-support-multi-vector.patch WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/external/faiss ERROR_VARIABLE ERROR_MSG RESULT_VARIABLE RESULT_CODE)
     execute_process(COMMAND git ${GIT_PATCH_COMMAND} --3way --ignore-space-change --ignore-whitespace ${CMAKE_CURRENT_SOURCE_DIR}/patches/faiss/0002-Enable-precomp-table-to-be-shared-ivfpq.patch WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/external/faiss ERROR_VARIABLE ERROR_MSG RESULT_VARIABLE RESULT_CODE)

jni/cmake/init-nmslib.cmake

@@ -12,11 +12,25 @@ if (NOT EXISTS ${NMS_REPO_DIR})
     execute_process(COMMAND git submodule update --init -- external/nmslib WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR})
 endif ()
 
-# Check if patch exist, this is to skip git apply during CI build. See CI.yml with ubuntu.
-find_path(PATCH_FILE NAMES 0001-Initialize-maxlevel-during-add-from-enterpoint-level.patch 0002-Adds-ability-to-pass-ef-parameter-in-the-query-for-h.patch PATHS ${CMAKE_CURRENT_SOURCE_DIR}/patches/nmslib NO_DEFAULT_PATH)
+# Get the last patch file name
+file(GLOB PATCH_FILE_LIST "${CMAKE_CURRENT_SOURCE_DIR}/patches/nmslib/*")
+list(SORT PATCH_FILE_LIST ORDER DESCENDING)
+list(GET PATCH_FILE_LIST 0 LAST_PATCH_FILE)
+message(STATUS "Last patch file: ${LAST_PATCH_FILE}")
 
-# If it exists, apply patches
-if (EXISTS ${PATCH_FILE})
+# Get patch id of the last patch
+execute_process(COMMAND sh -c "cat ${LAST_PATCH_FILE} | git patch-id --stable" OUTPUT_VARIABLE PATCH_ID_OUTPUT)
+string(REPLACE " " ";" PATCH_ID_LIST ${PATCH_ID_OUTPUT})
+list(GET PATCH_ID_LIST 0 PATCH_ID)
+message(STATUS "Extracted patch ID: ${PATCH_ID}")
+
+# Get patch id of the last commit
+execute_process(COMMAND sh -c "git --no-pager show HEAD | git patch-id --stable" OUTPUT_VARIABLE PATCH_ID_OUTPUT_FROM_COMMIT WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/external/nmslib)
+string(REPLACE " " ";" PATCH_ID_LIST_FROM_COMMIT ${PATCH_ID_OUTPUT_FROM_COMMIT})
+list(GET PATCH_ID_LIST_FROM_COMMIT 0 PATCH_ID_FROM_COMMIT)
+message(STATUS "Extracted patch ID from commit: ${PATCH_ID_FROM_COMMIT}")
+
+if (NOT ${PATCH_ID} STREQUAL ${PATCH_ID_FROM_COMMIT})
     message(STATUS "Applying custom patches.")
     execute_process(COMMAND git ${GIT_PATCH_COMMAND} --3way --ignore-space-change --ignore-whitespace ${CMAKE_CURRENT_SOURCE_DIR}/patches/nmslib/0001-Initialize-maxlevel-during-add-from-enterpoint-level.patch WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/external/nmslib ERROR_VARIABLE ERROR_MSG RESULT_VARIABLE RESULT_CODE)
     execute_process(COMMAND git ${GIT_PATCH_COMMAND} --3way --ignore-space-change --ignore-whitespace ${CMAKE_CURRENT_SOURCE_DIR}/patches/nmslib/0002-Adds-ability-to-pass-ef-parameter-in-the-query-for-h.patch WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/external/nmslib ERROR_VARIABLE ERROR_MSG RESULT_VARIABLE RESULT_CODE)