diff --git a/how-tos/compute-to-data/make-a-boss-c2d-algorithm.md b/how-tos/compute-to-data/make-a-boss-c2d-algorithm.md index c5ebc26c..be82c515 100644 --- a/how-tos/compute-to-data/make-a-boss-c2d-algorithm.md +++ b/how-tos/compute-to-data/make-a-boss-c2d-algorithm.md @@ -6,6 +6,8 @@ description: >- # Make a Boss C2D Algorithm +{% embed url="https://media3.giphy.com/media/cub3pntkz8muQ/giphy.gif?cid=ecf05e47o9fdrqco4jqpeyh7899whqgw5tnd43elr023rykr&ep=v1_gifs_search&rid=giphy.gif" %} + The beginning of any great algorithm for Compute-to-Data first starts by referencing the dataset file correctly on the Docker image. Here is the code in Javascript and in Python for how to correctly reference the dataset file that's local on the Docker container that you choose: ### Python @@ -41,6 +43,8 @@ if not input_filename: with open(input_filename, 'r') as file: # Read the CSV file csv_reader = csv.DictReader(file) + + ``` ### Javascript