Python - 如何通过循环分页 API 来提取数据（Harvest）

Question

First of all I have been working with Python for about a couple of days, so I don't necessarily know the best practices or all the terminology ... yet. I learn best by reverse engineering and my code below is based on the official documentation from Harvest and other bits I've found with google-fu

My request is to download all the time entries records from Harvest and save as a JSON (or ideally a CSV file).

Official Python Example from Harvest Git Hub

This is my adapted code (including all outputs, which won't be neccessary in the final code but handy for my learning):

import requests, json, urllib.request

#Set variables for authorisation
AUTH = "REDACTED"
ACCOUNT = "REDACTED"

URL = "https://api.harvestapp.com/v2/time_entries"
HEADERS = { "Authorization": AUTH,
            "Harvest-Account-ID": ACCOUNT}
PAGENO = str("5")

request = urllib.request.Request(url=URL+"?page="+PAGENO, headers=HEADERS)
response = urllib.request.urlopen(request, timeout=5)
responseBody = response.read().decode("utf-8")
jsonResponse = json.loads(responseBody)

# Find the values for pagination
parsed = json.loads(responseBody)
links_first = parsed["links"]["first"]
links_last = parsed["links"]["last"]
links_next = parsed["links"]["next"]
links_previous = parsed["links"]["previous"]
nextpage = parsed["next_page"]
page = parsed["page"]
perpage = parsed["per_page"]
prevpage = parsed["previous_page"]
totalentries = parsed["total_entries"]
totalpages = parsed["total_pages"]

#Print the output
print(json.dumps(jsonResponse, sort_keys=True, indent=4))
print("first link : " + links_first)
print("last link : " + links_last)
print("next page : " + str(nextpage))
print("page : " + str(page))
print("per page : " + str(perpage))
print("total records : " + str(totalentries))
print("total pages : " + str(totalpages))

The output response is
"Squeezed text (5816 lines)"
first link : https://api.harvestapp.com/v2/time_entries?page=1&per_page=100
last link : https://api.harvestapp.com/v2/time_entries?page=379&per_page=100
next page : 6
page : 5
per page : 100
total records : 37874
total pages : 379

Please can someone advise the best way to loop through the pages to form one JSON file ? If you are also able to advise the best way then output that JSON file I would be very grateful.

Answer 1

I have been using the following code to retrieve all time entries. It could be a bit more effective, perhaps, but it works. The function get_all_time_entries loops through all the pages and appends the response in JSON format into all_time_entries array and finally returns this array.

import requests
import json

def get_all_time_entries():

    url_address = "https://api.harvestapp.com/v2/time_entries"  
    headers = {
        "Authorization": "Bearer " + "xxxxxxxxxx",
        "Harvest-Account-ID": "xxxxxx"
    }

    # find out total number of pages
    r = requests.get(url=url_address, headers=headers).json()
    total_pages = int(r['total_pages'])

    # results will be appended to this list
    all_time_entries = []

    # loop through all pages and return JSON object
    for page in range(1, total_pages):

        url = "https://api.harvestapp.com/v2/time_entries?page="+str(page)              
        response = requests.get(url=url, headers=headers).json()        
        all_time_entries.append(response)       
        page += 1

    # prettify JSON
    data = json.dumps(all_time_entries, sort_keys=True, indent=4)

    return data

print(get_all_time_entries())

You can easily direct the output of the script with ">" to local folder when running in powershell, etc.

For example:

Python.exe example.py > C:\\temp\\all_time_entries.json

Hope this helps!

Answer 2

There's a Python library that supports Harvest API v2.

The library supports all of the authentication methods, request rate limiting, response codes and has dataclasses for each of the response objects.

The library is very well tested so you will have an example of usage for each endpoint in the tests. The tests use the official Harvest examples.

Additionally there is an example detailed time report which inherits the Harvest object. The tests for the detailed time report show how to use it.

The library is referenced from the Harvest software directory; https://www.getharvest.com/integrations/python-library

Project URL; https://github.com/bradbase/python-harvest_apiv2

I own the project.

from harvest import Harvest
from .harvestdataclasses import *

class MyTimeEntries(Harvest):

    def __init__(self, uri, auth):
        super().__init__(uri, auth)


    def time_entries(self):
        time_entry_results = []
       
        time_entries = self.time_entries()
        time_entry_results.extend(time_entries.time_entries)
        if time_entries.total_pages > 1:
            for page in range(2, time_entries.total_pages + 1):
                time_entries = self.time_entries(page=page)
                time_entry_results.extend(time_entries.time_entries)
        
        return time_entry_results

personal_access_token = PersonalAccessToken('ACCOUNT_NUMBER', 'PERSONAL_ACCESS_TOKEN')
my_report = MyTimeEntries('https://api.harvestapp.com/api/v2', personal_access_token)
time_entries = my_report.time_entries()