finalproj_upload.py

#
# Uploads a PDF to S3 and then inserts a new job record
# in the Transcripto database with a status of 'uploaded'.
# Sends the job id back to the client.
#

import json
import boto3
import os
import uuid
import base64
import pathlib
import datatier
import tempfile

from configparser import ConfigParser

def lambda_handler(event, context):
  try:
    print("**STARTING**")
    print("**lambda: final proj upload**")
    
    #
    # setup AWS based on config file:
    #
    config_file = 'transcripto-app-config.ini'
    os.environ['AWS_SHARED_CREDENTIALS_FILE'] = config_file
    
    configur = ConfigParser()
    configur.read(config_file)
    
    #
    # configure for S3 access:
    # #
    # s3_profile = 's3readwrite'
    # boto3.setup_default_session(profile_name=s3_profile)
    
    bucketname = configur.get('s3', 'bucket_name')
    
    s3 = boto3.resource('s3')
    bucket = s3.Bucket(bucketname)
    
    #
    # configure for RDS access
    #
    rds_endpoint = configur.get('rds', 'endpoint')
    rds_portnum = int(configur.get('rds', 'port_number'))
    rds_username = configur.get('rds', 'user_name')
    rds_pwd = configur.get('rds', 'user_pwd')
    rds_dbname = configur.get('rds', 'db_name')
    
  
    #
    # the user has sent us two parameters:
    #  1. filename of their file
    #  2. raw file data in base64 encoded string
    #
    # The parameters are coming through web server 
    # (or API Gateway) in the body of the request
    # in JSON format.
    #
    print("**Accessing request body**")

    if "body" not in event:
        raise Exception("event has no body")

    try:
        body = json.loads(event["body"])
        filename = body["filename"]
        datastr = body["data"]
        file_bytes = base64.b64decode(datastr)
    except Exception as e:
        return {
            "statusCode": 400,
            "body": json.dumps({"error": "Invalid request body: must contain base64-encoded 'data' and 'filename'"})
        }

    # open connection to the database:
    #
    print("**Opening connection**")
    
    dbConn = datatier.get_dbConn(rds_endpoint, rds_portnum, rds_username, rds_pwd, rds_dbname)

    print("**dbconn worked.")
    #
    # upload to s3
    #
    base64_bytes = datastr.encode()        # string -> base64 bytes
    bytes = base64.b64decode(base64_bytes) # base64 bytes -> raw bytes
    
    #
    # write raw bytes to local filesystem for upload:
    #
    print("**Writing local data file**")
    #
    # TODO #1 of 3: what directory do we write to locally?
    # Then open this local file for writing a binary file,
    # write the bytes we received from the client, and
    # close the file.
    #
    # headers = event.get("headers", {})
    # content_type = headers.get("Content-Type", headers.get("content-type", ""))
    # print("Content-Type:", content_type)

    # if not content_type.startswith("multipart/form-data"):
    #     return {
    #         "statusCode": 400,
    #         "body": json.dumps({"error": "Invalid request. Must be multipart/form-data."})
    #     }    
    # Write file to temporary storage
    
    temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=pathlib.Path(filename).suffix)
    temp_file.write(file_bytes)
    temp_file.close()
    local_filename = temp_file.name

    
    print(f"Saved uploaded file to: {temp_file.name}")
    
    #
    # generate unique filename in preparation for the S3 upload:
    #
    print("**Uploading local file to S3**")
    
    basename = pathlib.Path(filename).stem
    extension = pathlib.Path(filename).suffix
        
    bucketkey = "transcriptoapp/" + basename + "-" + str(uuid.uuid4()) + extension

    print("S3 bucketkey:", bucketkey)

    query_params = event.get("queryStringParameters", {})
    job_type = query_params.get("job_type", "").lower()
        
    if job_type not in {"transcription", "text_to_speech", "translation"}:
        return {
            "statusCode": 400,
            "body": json.dumps({"error": "Invalid job_type. Must be 'transcription', 'text_to_speech', or 'translation'."})
        }

    #
    # Remember that the processing of the PDF is event-triggered,
    # and that lambda function is going to update the database as
    # is processes. So let's insert a job record into the database
    # first, THEN upload the PDF to S3. The status column should 
    # be set to 'uploaded':
    #
    print("**Adding jobs row to database**")
        
    sql = """
        INSERT INTO jobs (jobtype, status, originaldatafile, datafilekey, resultsfilekey)
        VALUES (%s, 'uploaded', %s, %s, '');
    """    
    #
    # TODO #2 of 3: what values should we insert into the database?
    #
    datatier.perform_action(dbConn, sql, [job_type, filename, bucketkey])

    #
    # grab the jobid that was auto-generated by mysql:
    #
    sql = "SELECT LAST_INSERT_ID();"
    
    row = datatier.retrieve_one_row(dbConn, sql)
    
    jobid = row[0]
    
    print("jobid:", jobid)

    #
    # now that DB is updated, let's upload PDF to S3:
    #
    print("**Uploading data file to S3**")

    #
    # TODO #3 of 3: what are we uploading to S3? replace the
    # ??? with what we are uploading:
    #
    bucket.upload_file(
        local_filename,
        bucketkey,
        ExtraArgs={'ACL': 'public-read', 'ContentType': 'application/octet-stream'}
    )

    #
    # respond in an HTTP-like way, i.e. with a status
    # code and body in JSON format:
    #
    print("**DONE, returning jobid**")
    
    return {
      'statusCode': 200,
      'body': json.dumps(str(jobid))
    }
    
  except Exception as err:
    print("**ERROR**")
    print(str(err))
    
    return {
      'statusCode': 500,
      'body': json.dumps(str(err))
    }