Skip to main content
 首页 » 编程设计

python之Facebook 广告 API之批量请求定向搜索

2025年02月15日7over140


问题:

我在为 Facebook 广告 API 提交批量请求时遇到问题,我想知道是否有人可以提供有关以下错误的见解。

我试图获取一个艺术家列表,在这个简化示例中只有 50 个,然后为这些艺术家提交 TargetingSearch 请求。但是,我无法将请求正确传递给 Facebook。您可以想象我将成千上万的艺术家传递到这里的情况。

我一直在使用这两个资源在 facebook-ads-api for Python 中批量请求广告/广告集洞察作为引用: https://github.com/facebook/facebook-python-ads-sdk/issues/116 https://github.com/linpingta/facebook-related/blob/master/facebook-ads-sdk-example/insight_related.py

我的代码:

import time 
list_of_artists = ['Adele','Alessia Cara','Ariana Grande','Big Sean','Blake Shelton','Brantley Gilbert','Brett Young','Bruno Mars','Calvin Harris','Camila Cabello','Carrie Underwood','Chris Brown','Chris Stapleton','Chuck Berry','Cole Swindell','Dierks Bentley','DJ Khaled','Ed Sheeran','Eminem','Eric Church','Future','G-Eazy','Gucci Mane','James Arthur','Jason Aldean','John Legend','Jon Pardi','Josh Turner','Julia Michaels','Justin Bieber','Justin Timberlake','Katy Perry','Kehlani','Keith Urban','Kelsea Ballerini','Kendrick Lamar','Kenny Chesney','Khalid','Kodak Black','Kygo','Kyle Harvey','Lady Gaga','Lil Yachty','Lorde','Luis Fonsi','Luke Bryan','Luke Combs','Machine Gun Kelly (Rapper)'] 
 
def success_callback(response): 
    batch_body_responses.append(response.body()) 
 
def error_callback(response): 
    # Error handling here 
    pass 
 
def get_id_list(art_search_list): 
    batches = [] 
    batch_body_responses = [] 
    your_app_id = '<appid>' 
    your_app_secret = '<appsecret>' 
    your_access_token = '<token>' 
    api = FacebookAdsApi.init(your_app_id, your_app_secret, your_access_token) 
    batch_limit = 25 
    for batch in generate_batches(art_search_list, batch_limit): 
        next_batch = api.new_batch() 
        for art in batch: 
            requests = TargetingSearch.search(params = {'q': art,'type': 'adinterest'}) 
            for req in requests: 
                # adding a print to see what req looks like 
                print req 
                next_batch.add_request(req, success_callback, error_callback) 
        batches.append(next_batch) 
 
    for batch_request in batches: 
        batch_request.execute() 
        time.sleep(5) 
 
print batch_body_responses 
 
get_id_list(list_of_artists) 

这会导致以下错误:

注意:TargetingSearch 来自上面的打印请求。

<TargetingSearch> { 
    "audience_size": 30176300, 
    "id": "6003173089178", 
    "name": "Adele", 
    "path": [ 
        "Interests", 
        "Additional Interests", 
        "Adele" 
    ], 
    "topic": "People" 
} 
--------------------------------------------------------------------------- 
AttributeError                            Traceback (most recent call last) 
<ipython-input-299-67aa05ef2228> in <module>() 
     33     print batch_body_responses 
     34  
---> 35 get_id_list(list_of_artists) 
     36  
     37 # # ################# 
 
<ipython-input-299-67aa05ef2228> in get_id_list(art_search_list) 
     24                 # adding a print to see what req looks like 
     25                 print req 
---> 26                 next_batch.add_request(req, success_callback, error_callback) 
     27         batches.append(next_batch) 
     28  
 
//anaconda/lib/python2.7/site-packages/facebookads/api.pyc in add_request(self, request, success, failure) 
    445                 A dictionary describing the call. 
    446         """ 
--> 447         updated_params = copy.deepcopy(request._params) 
    448         if request._fields: 
    449             updated_params['fields'] = ','.join(request._fields) 
 
AttributeError: 'TargetingSearch' object has no attribute '_params' 

如果我能正确地将批处理传递给 Facebook,我希望得到一个答案,为我提供每个艺术家的 TargetingSearch 结果。这是阿黛尔的示例,但您可以想象为每个人复制此示例:

[<TargetingSearch> { 
    "audience_size": 30176300, 
    "id": "6003173089178", 
    "name": "Adele", 
    "path": [ 
        "Interests", 
        "Additional Interests", 
        "Adele" 
    ], 
    "topic": "People" 
}, <TargetingSearch> { 
    "audience_size": 20449710, 
    "id": "6003701797690", 
    "name": "21 (Adele album)", 
    "path": [ 
        "Interests", 
        "Additional Interests", 
        "21 (Adele album)" 
    ], 
    "topic": "News and entertainment" 
}, <TargetingSearch> { 
    "audience_size": 256080, 
    "disambiguation_category": "Song", 
    "id": "6005916496872", 
    "name": "Someone like You (Adele song)", 
    "path": [ 
        "Interests", 
        "Additional Interests", 
        "Someone like You (Adele song)" 
    ], 
    "topic": "News and entertainment" 
}, <TargetingSearch> { 
    "audience_size": 130230, 
    "disambiguation_category": "Home", 
    "id": "6002994499323", 
    "name": "Adele Live", 
    "path": [ 
        "Interests", 
        "Additional Interests", 
        "Adele Live" 
    ] 
}, <TargetingSearch> { 
    "audience_size": 31410, 
    "disambiguation_category": "Song", 
    "id": "6004484416351", 
    "name": "Rumour Has It (Adele song)", 
    "path": [ 
        "Interests", 
        "Additional Interests", 
        "Rumour Has It (Adele song)" 
    ], 
    "topic": "News and entertainment" 
}, <TargetingSearch> { 
    "audience_size": 4560, 
    "disambiguation_category": "Public Figure", 
    "id": "6003446313480", 
    "name": "Adele Parks", 
    "path": [ 
        "Interests", 
        "Additional Interests", 
        "Adele Parks" 
    ], 
    "topic": "People" 
}, <TargetingSearch> { 
    "audience_size": 2290, 
    "id": "6002970343768", 
    "name": "Adele ring", 
    "path": [ 
        "Interests", 
        "Additional Interests", 
        "Adele ring" 
    ], 
    "topic": "Education" 
}, <TargetingSearch> { 
    "audience_size": 990, 
    "disambiguation_category": "Musician/Band", 
    "id": "6003213600533", 
    "name": "Adele - Official Website", 
    "path": [ 
        "Interests", 
        "Additional Interests", 
        "Adele - Official Website" 
    ] 
}] 

我想我在理解如何传递/自定义请求并在可以执行的批处理中生成它们时遇到了麻烦。

-谢谢, 约旦

请您参考如下方法:


我的解决方案(可能还有改进空间)

我最终想出了一个解决我的问题的方法,它很可能不是最好的,但它现在有效。我将我最初提交的问题和不正确的代码留在了下面自己的标题下,以防其他人了解使用 Facebook API 进行批处理请求并发现我如何解决此问题很有帮助。

我最初的担忧是正确的,我实际上是在尝试将 TargetingSearch 调用作为请求提交,但实际上它们在作为批处理请求传递之前就已执行。

我需要做的是:

  • 使用 FacebookRequest 通过“/targetingsearch”端点创建“GET”请求,然后使用 .add_params 附加查询参数,这些参数根据批处理中的项目而变化。

  • 在回调函数中添加一些错误处理。某些搜索的艺术家实际上可能不会返回任何结果,如果您尝试使用索引提取 ID 和名称,则会出现错误。

  • 我选择导入 pandas 并将结果转换为数据框,以便于阅读。

提示:故意向 FB 提交不正确的调用有助于我进行故障排除,因为吐回的错误消息会向您显示正在发出的实际请求。

代码:

import time 
from facebookads.api import FacebookRequest 
import pandas as pd 
account = '<act_accountID>' 
list_of_artists = ['Adele','Alessia Cara','Ariana Grande','Big Sean','Blake Shelton','Brantley Gilbert','Brett Young','Bruno Mars','Calvin Harris','Camila Cabello','Carrie Underwood','Chris Brown','Chris Stapleton','Chuck Berry','Cole Swindell','Dierks Bentley','DJ Khaled','Ed Sheeran','Eminem','Eric Church','Future','G-Eazy','Gucci Mane','James Arthur','Jason Aldean','John Legend','Jon Pardi','Josh Turner','Julia Michaels','Justin Bieber','Justin Timberlake','Katy Perry','Kehlani','Keith Urban','Kelsea Ballerini','Kendrick Lamar','Kenny Chesney','Khalid','Kodak Black','Kygo','Kyle Harvey','Lady Gaga','Lil Yachty','Lorde','Luis Fonsi','Luke Bryan','Luke Combs','Machine Gun Kelly (Rapper)'] 
 
batch_body_responses = [] 
 
def success_callback(response): 
    try: 
        pair = [response.json()["data"][0]["name"],response.json()["data"][0]["id"]] 
        batch_body_responses.append(pair) 
    except IndexError: 
        pass 
    except UnicodeEncodeError: 
        pass 
 
def error_callback(response): 
    pass 
 
def get_id_list(art_search_list): 
    batches = [] 
    your_app_id = '<appid>' 
    your_app_secret = '<appsecret>' 
    your_access_token = '<token>' 
    api = FacebookAdsApi.init(your_app_id, your_app_secret, your_access_token) 
    batch_limit = 25 
    for batch in generate_batches(art_search_list, batch_limit): 
        next_batch = api.new_batch() 
        for artt in batch: 
            requests = [FacebookRequest(node_id=account,method="GET",endpoint="/targetingsearch").add_params(params = {'q': artt,'type': 'adinterest'})] 
            for req in requests: 
                next_batch.add_request(req, success_callback, error_callback) 
        batches.append(next_batch) 
 
    for batch_request in batches: 
        batch_request.execute() 
        time.sleep(2) 
 
    return batch_body_responses 
 
df = pd.DataFrame(get_id_list(list_of_artists)) 
df 

结果:

0   Adele   6003173089178 
1   Alessia Cara    931161786931155 
2   Ariana Grande   6003032339492 
3   Big Sean    6002839083879 
4   Blake Shelton   6003145416068 
5   Brantley Gilbert    6003087234070 
6   Young & hip 6006520994425 
7   Bruno Mars  6003392437554 
8   Calvin Harris   6003700476383 
9   Fifth Harmony (Camila Cabello)  173362709468415 
10  Carrie Underwood    6003320610294 
11  Chris Brown 6003412995205 
12  Chris Stapleton 6003632267583 
13  Chuck Berry 6003381796604 
14  Cole Swindell   6015941143427 
15  Dierks Bentley  6003273151043 
16  DJ Khaled   6003664971094 
17  Ed Sheeran  6003704170491 
18  Eminem  6003135347608 
19  Eric Church 6003174532035 
20  Future  6003289081451 
21  G-Eazy  6006359491399 
22  Gucci Mane  6003291602130 
23  james arthur    6011082407659 
24  Jason Aldean    6003138841739 
25  John Legend 6003287333056 
26  Jon Pardi   6005274491737 
27  Josh Turner 6003547807427 
28  Justin Bieber   6003143596040 
29  Justin Timberlake   6003125864388 
30  Katy Perry  6003514925642 
31  Kehlani 1673020586319299 
32  Keith Urban 6003280487023 
33  Kelsea Ballerini    1649934611908779 
34  Kendrick Lamar  6003436621883 
35  Kenny Chesney   6003183761012 
36  Khalid Yasin    6003043109915 
37  Kygo    342546012613588 
38  List of American Horror Story characters    6014211362622 
39  Lady Gaga   6003144466384 
40  Lil Yachty  1177170802403085 
41  Lorde   6018616059273 
42  Luis Fonsi  6003210594524 
43  Luke Bryan  6003290279056 
44  Machine Gun Kelly (rapper)  6003461497689