Clean up CCI cache handling (#1238)

1. Add `generate_cachekey` and `fetch_cachekey` command that abstracts away the way cache key is generated.
2. Fix the daily cache key. `date +%D` (the previous `"+%Y-%d"` does not change on daily basis)
3. Remove conda environment cache for simplicity (this is no longer the bottleneck and having cached conda env often causes more trouble when macOS test is enabled.)

Author: mthrok

.circleci/config.yml

@@ -39,6 +39,21 @@ commands:
               if [[ "${PYTHON_VERSION}" = *3.9* ]]; then
                 echo "export CONDA_CHANNEL_FLAGS=-c=conda-forge" >> ${BASH_ENV}
               fi
+  generate_cachekey:
+    description: "Generate .cachekey file that changes on daily basis"
+    steps:
+      - run:
+          name: Generate CCI cache key
+          command: echo "$(date "+%D")" > .cachekey
+      - persist_to_workspace:
+          root: .
+          paths:
+            - .cachekey
+  fetch_cachekey:
+    description: "Fetch the .cachekey file that is generated by generate_cachekey job"
+    steps:
+      - attach_workspace:
+          at: .
 
 binary_common: &binary_common
   parameters:
@@ -356,64 +371,34 @@ jobs:
       - checkout
       - designate_upload_channel
       - load_conda_channel_flags
-      - run:
-          name: Generate cache key
-          # This will refresh cache on Sundays, nightly build should generate new cache.
-          command: |
-            echo "$(date +"Year-Week %Y-%U")" > .circleci-weekly
-            echo "$(date +"Year-Date %Y-%d")" > .circleci-daily-linux
-      - persist_to_workspace:
-          root: .
-          paths:
-            - .circleci-daily-linux
+      - generate_cachekey
       - restore_cache:
           keys:
 
-            - data-linux-v1-{{ checksum ".circleci-daily-linux" }}
+            - v1-linux-cache-index-{{ checksum ".cachekey" }}
 
       - run:
-          name: Exit if cache exists
+          name: Generate cache
+          no_output_timeout: 30m
           command: |
-            FILE=.data/cache_status_file.json
-            if test -f "$FILE"; then
-              cat $FILE
-              circleci step halt
-            else
-              echo "$FILE does not exists"
+            if [ ! -f .data/cache_status_file.json ] ; then
+              .circleci/unittest/linux/scripts/setup_env.sh
+              .circleci/unittest/linux/scripts/install.sh
+              .circleci/unittest/linux/scripts/generate_cache.sh
             fi
-      - restore_cache:
-
-          keys:
-            - env-v1-linux-{{ arch }}-py<< parameters.python_version >>-{{ checksum ".circleci/unittest/linux/scripts/environment.yml" }}-{{ checksum ".circleci-weekly" }}
-
-      - run:
-          name: Setup
-          command: .circleci/unittest/linux/scripts/setup_env.sh
+            cat .data/cache_status_file.json
       - save_cache:
 
-          key: env-v1-linux-{{ arch }}-py<< parameters.python_version >>-{{ checksum ".circleci/unittest/linux/scripts/environment.yml" }}-{{ checksum ".circleci-weekly" }}
+          key: v1-linux-dataset-{{ checksum ".cachekey" }}
 
           paths:
-            - conda
-            - env
-      - run:
-          name: Install torchtext
-          command: .circleci/unittest/linux/scripts/install.sh
-      - run:
-          name: Generate daily data Cache
-          no_output_timeout: 30m
-          command: |
-            .circleci/unittest/linux/scripts/generate_cache.sh
-            cat .data/cache_status_file.json
+            - .data
       - save_cache:
 
-          key: data-linux-v1-{{ checksum ".circleci-daily-linux" }}
+          key: v1-linux-cache-index-{{ checksum ".cachekey" }}
 
           paths:
-            - .data
-      - run:
-          name: Post process
-          command: .circleci/unittest/linux/scripts/post_process.sh
+            - .data/cache_status_file.json
 
   unittest_linux:
     <<: *binary_common
@@ -424,40 +409,18 @@ jobs:
       - checkout
       - designate_upload_channel
       - load_conda_channel_flags
-      - attach_workspace:
-          at: .
-      - run:
-          name: Generate cache key
-          # This will refresh cache on Sundays, nightly build should generate new cache.
-          command: |
-            echo "$(date +"Year-Week %Y-%U")" > .circleci-weekly
-      - restore_cache:
-
-          keys:
-            - env-v1-linux-{{ arch }}-py<< parameters.python_version >>-{{ checksum ".circleci/unittest/linux/scripts/environment.yml" }}-{{ checksum ".circleci-weekly" }}
-
+      - fetch_cachekey
       - run:
           name: Setup
           command: .circleci/unittest/linux/scripts/setup_env.sh
-      - save_cache:
-
-          key: env-v1-linux-{{ arch }}-py<< parameters.python_version >>-{{ checksum ".circleci/unittest/linux/scripts/environment.yml" }}-{{ checksum ".circleci-weekly" }}
-
-          paths:
-            - conda
-            - env
       - run:
           name: Install torchtext
           command: .circleci/unittest/linux/scripts/install.sh
       - restore_cache:
           keys:
 
-            - data-linux-v1-{{ checksum ".circleci-daily-linux" }}
-
-      - restore_cache:
-          keys:
-
-            - data-vector-linux-v1-{{ checksum ".circleci-daily-linux" }}
+            - v1-linux-dataset-vector-{{ checksum ".cachekey" }}
+            - v1-linux-dataset-{{ checksum ".cachekey" }}
 
       - run:
           name: Run tests
@@ -466,7 +429,7 @@ jobs:
           command: .circleci/unittest/linux/scripts/run_test.sh
       - save_cache:
 
-          key: data-vector-linux-v1-{{ checksum ".circleci-daily-linux" }}
+          key: v1-linux-dataset-vector-{{ checksum ".cachekey" }}
 
           paths:
             - .vector_cache
@@ -485,67 +448,34 @@ jobs:
       - checkout
       - designate_upload_channel
       - load_conda_channel_flags
-      - run:
-          name: Generate cache key
-          # This will refresh cache on Sundays, nightly build should generate new cache.
-          command: |
-            echo "$(date +"Year-Week %Y-%U")" > .circleci-weekly
-            echo "$(date +"Year-Date %Y-%d")" > .circleci-daily-win
-      - persist_to_workspace:
-          root: .
-          paths:
-            - .circleci-daily-win
+      - generate_cachekey
       - restore_cache:
           keys:
 
-            - data-windows-v1-{{ checksum ".circleci-daily-win" }}
-
-      - run:
-          name: Exit if cache exists
-          command: |
-            FILE=.data/cache_status_file.json
-            if test -f "$FILE"; then
-              cat $FILE
-              circleci step halt
-            else
-              echo "$FILE does not exists"
-            fi
-      - restore_cache:
-
-          keys:
-            - env-v1-windows-{{ arch }}-py<< parameters.python_version >>-{{ checksum ".circleci/unittest/windows/scripts/environment.yml" }}-{{ checksum ".circleci-weekly" }}
-
-      - run:
-          name: Setup
-          command: .circleci/unittest/windows/scripts/setup_env.sh
-      - save_cache:
-
-          key: env-v1-windows-{{ arch }}-py<< parameters.python_version >>-{{ checksum ".circleci/unittest/windows/scripts/environment.yml" }}-{{ checksum ".circleci-weekly" }}
+            - v1-windows-cache-index-{{ checksum ".cachekey" }}
 
-          paths:
-            - conda
-            - env
-      - run:
-          name: Install torchtext
-          command: .circleci/unittest/windows/scripts/install.sh
       - run:
           name: Generate daily data Cache
           no_output_timeout: 30m
           command: |
-            .circleci/unittest/windows/scripts/generate_cache.sh
+            if [ ! -f .data/cache_status_file.json ] ; then
+              .circleci/unittest/windows/scripts/setup_env.sh
+              .circleci/unittest/windows/scripts/install.sh
+              .circleci/unittest/windows/scripts/generate_cache.sh
+            fi
             cat .data/cache_status_file.json
       - save_cache:
 
-          key: data-windows-v1-{{ checksum ".circleci-daily-win" }}
+          key: v1-windows-dataset-{{ checksum ".cachekey" }}
 
           paths:
             - .data
-      - run:
-          name: Post process
-          command: .circleci/unittest/windows/scripts/post_process.sh
-
+      - save_cache:
 
+          key: v1-windows-cache-index-{{ checksum ".cachekey" }}
 
+          paths:
+            - .data/cache_status_file.json
 
   unittest_windows:
     <<: *binary_common
@@ -555,40 +485,18 @@ jobs:
       - checkout
       - designate_upload_channel
       - load_conda_channel_flags
-      - attach_workspace:
-          at: .
-      - run:
-          name: Generate cache key
-          # This will refresh cache on Sundays, nightly build should generate new cache.
-          command: |
-            echo "$(date +"Year-Week %Y-%U")" > .circleci-weekly
-      - restore_cache:
-
-          keys:
-            - env-v1-windows-{{ arch }}-py<< parameters.python_version >>-{{ checksum ".circleci/unittest/windows/scripts/environment.yml" }}-{{ checksum ".circleci-weekly" }}
-
+      - fetch_cachekey
       - run:
           name: Setup
           command: .circleci/unittest/windows/scripts/setup_env.sh
-      - save_cache:
-
-          key: env-v1-windows-{{ arch }}-py<< parameters.python_version >>-{{ checksum ".circleci/unittest/windows/scripts/environment.yml" }}-{{ checksum ".circleci-weekly" }}
-
-          paths:
-            - conda
-            - env
       - run:
           name: Install torchtext
           command: .circleci/unittest/windows/scripts/install.sh
       - restore_cache:
           keys:
 
-            - data-windows-v1-{{ checksum ".circleci-daily-win" }}
-
-      - restore_cache:
-          keys:
-
-            - data-vector-windows-v1-{{ checksum ".circleci-daily-win" }}
+            - v1-windows-dataset-vector-{{ checksum ".cachekey" }}
+            - v1-windows-dataset-{{ checksum ".cachekey" }}
 
 
       - run:
@@ -598,7 +506,7 @@ jobs:
           command: .circleci/unittest/windows/scripts/run_test.sh
       - save_cache:
 
-          key: data-vector-windows-v1-{{ checksum ".circleci-daily-win" }}
+          key: v1-windows-dataset-vector-{{ checksum ".cachekey" }}
 
           paths:
             - .vector_cache
@@ -617,25 +525,9 @@ jobs:
     steps:
       - checkout
       - designate_upload_channel
-      - run:
-          name: Generate cache key
-          # This will refresh cache on Sundays, nightly build should generate new cache.
-          command: echo "$(date +"Year-Week %Y-%U")" > .circleci-weekly
-      - restore_cache:
-
-          keys:
-            - env-v1-linux-{{ arch }}-py<< parameters.python_version >>-{{ checksum ".circleci/unittest/linux/scripts/environment.yml" }}-{{ checksum ".circleci-weekly" }}
-
       - run:
           name: Setup
           command: .circleci/unittest/linux/scripts/setup_env.sh
-      - save_cache:
-
-          key: env-v1-linux-{{ arch }}-py<< parameters.python_version >>-{{ checksum ".circleci/unittest/linux/scripts/environment.yml" }}-{{ checksum ".circleci-weekly" }}
-
-          paths:
-            - conda
-            - env
       - run:
           name: Run style check
           command: .circleci/unittest/linux/scripts/run_style_checks.sh