test if print changed files names work #22
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: Check consistency of tokens.txt file | |
| on: | |
| push: | |
| paths: | |
| - "chebai/preprocessing/bin/smiles_token/tokens.txt" | |
| - "chebai/preprocessing/bin/smiles_token_unlabeled/tokens.txt" | |
| - "chebai/preprocessing/bin/selfies/tokens.txt" | |
| - "chebai/preprocessing/bin/protein_token/tokens.txt" | |
| - "chebai/preprocessing/bin/graph_properties/tokens.txt" | |
| - "chebai/preprocessing/bin/graph/tokens.txt" | |
| - "chebai/preprocessing/bin/deepsmiles_token/tokens.txt" | |
| pull_request: | |
| paths: | |
| - "chebai/preprocessing/bin/smiles_token/tokens.txt" | |
| - "chebai/preprocessing/bin/smiles_token_unlabeled/tokens.txt" | |
| - "chebai/preprocessing/bin/selfies/tokens.txt" | |
| - "chebai/preprocessing/bin/protein_token/tokens.txt" | |
| - "chebai/preprocessing/bin/graph_properties/tokens.txt" | |
| - "chebai/preprocessing/bin/graph/tokens.txt" | |
| - "chebai/preprocessing/bin/deepsmiles_token/tokens.txt" | |
| jobs: | |
| check_tokens: | |
| runs-on: ubuntu-latest | |
| steps: | |
| - name: Checkout code | |
| uses: actions/checkout@v2 | |
| - name: Get list of changed files | |
| id: changed_files | |
| run: | | |
| # Get the list of changed files and save them to a file | |
| git diff --name-only origin/main > changed_files.txt | |
| # Print the names of changed files on separate lines | |
| echo "Changed files:" | |
| while read -r line; do | |
| echo "Changed File name : $line" | |
| done < changed_files.txt | |
| - name: Set global variable for tokens.txt path | |
| run: | | |
| TOKENS_FILE_PATH="chebai/preprocessing/bin/smiles_token/tokens.txt" | |
| echo "TOKENS_FILE_PATH=$TOKENS_FILE_PATH" >> $GITHUB_ENV | |
| - name: Get previous tokens.txt version | |
| run: | | |
| git fetch origin dev | |
| git diff origin/dev -- $TOKENS_FILE_PATH > tokens_diff.txt || echo "No previous tokens.txt found" | |
| - name: Check for deleted or added lines in tokens.txt | |
| run: | | |
| if [ -f tokens_diff.txt ]; then | |
| # Check for deleted lines (lines starting with '-') | |
| deleted_lines=$(grep '^-' tokens_diff.txt | grep -v '^---' | sed 's/^-//' || true) | |
| if [ -n "$deleted_lines" ]; then | |
| echo "Error: Lines have been deleted from tokens.txt. file" | |
| echo -e "Deleted Lines: \n$deleted_lines" | |
| exit 1 | |
| fi | |
| # Check for added lines (lines starting with '+') | |
| added_lines=$(grep '^+' tokens_diff.txt | grep -v '^+++' | sed 's/^+//' || true) | |
| if [ -n "$added_lines" ]; then | |
| # Count how many lines have been added | |
| num_added_lines=$(echo "$added_lines" | wc -l) | |
| # Get last `n` lines (equal to num_added_lines) of tokens.tx | |
| last_lines=$(tail -n "$num_added_lines" $TOKENS_FILE_PATH) | |
| # Check if the added lines are at the end of the file | |
| if [ "$added_lines" != "$last_lines" ]; then | |
| # Find lines that were added but not appended at the end of the file | |
| non_appended_lines=$(diff <(echo "$added_lines") <(echo "$last_lines") | grep '^<' | sed 's/^< //') | |
| echo "Error: New lines have been added, but they are not at the end of tokens.txt." | |
| echo -e "Added lines that are not at end of file: \n$non_appended_lines" | |
| exit 1 | |
| fi | |
| fi | |
| if [ "$added_lines" == "" ]; then | |
| echo "tokens.txt validation successful: No lines were deleted, and no new lines were added." | |
| else | |
| echo "tokens.txt validation successful: No lines were deleted, and new lines were correctly appended at the end." | |
| fi | |
| else | |
| echo "No previous version of tokens.txt found." | |
| fi |