跳转至

Search Engine Module

pipelines.pipelines.nodes.search_engine.providers

SearchApi

SearchApi is a real-time search engine that provides an API to access search results from Google, Google Scholar, YouTube, YouTube transcripts and more. See the SearchApi website for more details.

Source code in pipelines/pipelines/nodes/search_engine/providers.py 
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
class SearchApi(SearchEngine):
    """
    SearchApi is a real-time search engine that provides an API to access search results from Google, Google Scholar, YouTube,
    YouTube transcripts and more. See the [SearchApi website](https://www.searchapi.io/) for more details.
    """

    def __init__(
        self,
        api_key: str,
        top_k: Optional[int] = 10,
        engine: Optional[str] = "google",
        search_engine_kwargs: Optional[Dict[str, Any]] = None,
    ):
        """
        :param api_key: API key for SearchApi.
        :param top_k: Number of results to return.
        :param engine: Search engine to use, for example google, google_scholar, youtube, youtube_transcripts.
        See the [SearchApi documentation](https://www.searchapi.io/docs/google) for the full list of supported engines.
        :param search_engine_kwargs: Additional parameters passed to the SearchApi.
        See the [SearchApi documentation](https://www.searchapi.io/docs/google) for the full list of supported parameters.
        """
        super().__init__()
        self.params_dict: Dict[str, Union[str, int, float]] = {}
        self.api_key = api_key
        self.kwargs = search_engine_kwargs if search_engine_kwargs else {}
        self.engine = engine
        self.top_k = top_k

    def search(self, query: str, **kwargs) -> List[Document]:
        """
        :param query: Query string.
        :param kwargs: Additional parameters passed to the SearchApi. For example, you can set 'location' to 'New York,United States'
        to localize search to the specific location.
        :return: List[Document]
        """
        kwargs = {**self.kwargs, **kwargs}
        top_k = kwargs.pop("top_k", self.top_k)
        url = "https://www.searchapi.io/api/v1/search"

        params = {"q": query, **kwargs}
        headers = {"Authorization": f"Bearer {self.api_key}", "X-SearchApi-Source": "PaddleNLP"}

        if self.engine:
            params["engine"] = self.engine
        response = requests.get(url, params=params, headers=headers, timeout=90)

        if response.status_code != 200:
            raise Exception(f"Error while querying {self.__class__.__name__}: {response.text}")

        json_content = json.loads(response.text)
        documents = []
        has_answer_box = False

        if json_content.get("answer_box"):
            if json_content["answer_box"].get("organic_result"):
                title = json_content["answer_box"].get("organic_result").get("title", "")
                link = json_content["answer_box"].get("organic_result").get("link", "")
            if json_content["answer_box"].get("type") == "population_graph":
                title = json_content["answer_box"].get("place", "")
                link = json_content["answer_box"].get("explore_more_link", "")

            title = json_content["answer_box"].get("title", "")
            link = json_content["answer_box"].get("link")
            content = json_content["answer_box"].get("answer") or json_content["answer_box"].get("snippet")

            if link and content:
                has_answer_box = True
                documents.append(Document.from_dict({"title": title, "content": content, "link": link}))

        if json_content.get("knowledge_graph"):
            if json_content["knowledge_graph"].get("source"):
                link = json_content["knowledge_graph"].get("source").get("link", "")

            link = json_content["knowledge_graph"].get("website", "")
            content = json_content["knowledge_graph"].get("description")

            if link and content:
                documents.append(
                    Document.from_dict(
                        {"title": json_content["knowledge_graph"].get("title", ""), "content": content, "link": link}
                    )
                )

        documents += [
            Document.from_dict({"title": c["title"], "content": c.get("snippet", ""), "link": c["link"]})
            for c in json_content["organic_results"]
        ]

        if json_content.get("related_questions"):
            for question in json_content["related_questions"]:
                if question.get("source"):
                    link = question.get("source").get("link", "")
                else:
                    link = ""

                content = question.get("answer", "")

                if link and content:
                    documents.append(
                        Document.from_dict({"title": question.get("question", ""), "content": content, "link": link})
                    )

        logger.debug("SearchApi returned %s documents for the query '%s'", len(documents), query)
        result_docs = documents[:top_k]
        return self.score_results(result_docs, has_answer_box)

__init__ 

__init__(api_key: str, top_k: Optional[int] = 10, engine: Optional[str] = 'google', search_engine_kwargs: Optional[Dict[str, Any]] = None)

Parameters:

Name Type Description Default
api_key str

API key for SearchApi.

 required
top_k Optional[int]

Number of results to return.

 10
engine Optional[str]

Search engine to use, for example google, google_scholar, youtube, youtube_transcripts. See the SearchApi documentation for the full list of supported engines.

 'google'
search_engine_kwargs Optional[Dict[str, Any]]

Additional parameters passed to the SearchApi. See the SearchApi documentation for the full list of supported parameters.

 None
Source code in pipelines/pipelines/nodes/search_engine/providers.py 
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
def __init__(
    self,
    api_key: str,
    top_k: Optional[int] = 10,
    engine: Optional[str] = "google",
    search_engine_kwargs: Optional[Dict[str, Any]] = None,
):
    """
    :param api_key: API key for SearchApi.
    :param top_k: Number of results to return.
    :param engine: Search engine to use, for example google, google_scholar, youtube, youtube_transcripts.
    See the [SearchApi documentation](https://www.searchapi.io/docs/google) for the full list of supported engines.
    :param search_engine_kwargs: Additional parameters passed to the SearchApi.
    See the [SearchApi documentation](https://www.searchapi.io/docs/google) for the full list of supported parameters.
    """
    super().__init__()
    self.params_dict: Dict[str, Union[str, int, float]] = {}
    self.api_key = api_key
    self.kwargs = search_engine_kwargs if search_engine_kwargs else {}
    self.engine = engine
    self.top_k = top_k

search 

search(query: str, **kwargs) -> List[Document]

Parameters:

Name Type Description Default
query str

Query string.

 required
kwargs

Additional parameters passed to the SearchApi. For example, you can set 'location' to 'New York,United States' to localize search to the specific location.

 {}

Returns:

Type Description
List[Document]

List[Document]
Source code in pipelines/pipelines/nodes/search_engine/providers.py 
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
def search(self, query: str, **kwargs) -> List[Document]:
    """
    :param query: Query string.
    :param kwargs: Additional parameters passed to the SearchApi. For example, you can set 'location' to 'New York,United States'
    to localize search to the specific location.
    :return: List[Document]
    """
    kwargs = {**self.kwargs, **kwargs}
    top_k = kwargs.pop("top_k", self.top_k)
    url = "https://www.searchapi.io/api/v1/search"

    params = {"q": query, **kwargs}
    headers = {"Authorization": f"Bearer {self.api_key}", "X-SearchApi-Source": "PaddleNLP"}

    if self.engine:
        params["engine"] = self.engine
    response = requests.get(url, params=params, headers=headers, timeout=90)

    if response.status_code != 200:
        raise Exception(f"Error while querying {self.__class__.__name__}: {response.text}")

    json_content = json.loads(response.text)
    documents = []
    has_answer_box = False

    if json_content.get("answer_box"):
        if json_content["answer_box"].get("organic_result"):
            title = json_content["answer_box"].get("organic_result").get("title", "")
            link = json_content["answer_box"].get("organic_result").get("link", "")
        if json_content["answer_box"].get("type") == "population_graph":
            title = json_content["answer_box"].get("place", "")
            link = json_content["answer_box"].get("explore_more_link", "")

        title = json_content["answer_box"].get("title", "")
        link = json_content["answer_box"].get("link")
        content = json_content["answer_box"].get("answer") or json_content["answer_box"].get("snippet")

        if link and content:
            has_answer_box = True
            documents.append(Document.from_dict({"title": title, "content": content, "link": link}))

    if json_content.get("knowledge_graph"):
        if json_content["knowledge_graph"].get("source"):
            link = json_content["knowledge_graph"].get("source").get("link", "")

        link = json_content["knowledge_graph"].get("website", "")
        content = json_content["knowledge_graph"].get("description")

        if link and content:
            documents.append(
                Document.from_dict(
                    {"title": json_content["knowledge_graph"].get("title", ""), "content": content, "link": link}
                )
            )

    documents += [
        Document.from_dict({"title": c["title"], "content": c.get("snippet", ""), "link": c["link"]})
        for c in json_content["organic_results"]
    ]

    if json_content.get("related_questions"):
        for question in json_content["related_questions"]:
            if question.get("source"):
                link = question.get("source").get("link", "")
            else:
                link = ""

            content = question.get("answer", "")

            if link and content:
                documents.append(
                    Document.from_dict({"title": question.get("question", ""), "content": content, "link": link})
                )

    logger.debug("SearchApi returned %s documents for the query '%s'", len(documents), query)
    result_docs = documents[:top_k]
    return self.score_results(result_docs, has_answer_box)

SerpAPI

SerpAPI is a search engine that provides a REST API to access search results from Google, Bing, Yahoo, Yandex, Amazon, and similar. See the SerpAPI website for more details.

Source code in pipelines/pipelines/nodes/search_engine/providers.py 
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
class SerpAPI(SearchEngine):
    """
    SerpAPI is a search engine that provides a REST API to access search results from Google, Bing, Yahoo, Yandex,
    Amazon, and similar. See the [SerpAPI website](https://serpapi.com/) for more details.
    """

    def __init__(
        self,
        api_key: str,
        top_k: Optional[int] = 10,
        engine: Optional[str] = "google",
        search_engine_kwargs: Optional[Dict[str, Any]] = None,
    ):
        """
        :param api_key: API key for SerpAPI.
        :param top_k: Number of results to return.
        :param engine: Search engine to use, for example google, bing, baidu, duckduckgo, yahoo, yandex.
        See the [SerpAPI documentation](https://serpapi.com/search-api) for the full list of supported engines.
        :param search_engine_kwargs: Additional parameters passed to the SerperDev API. For example, you can set 'lr' to 'lang_en'
        to limit the search to English.
        See the [SerpAPI documentation](https://serpapi.com/search-api) for the full list of supported parameters.
        """
        super().__init__()
        self.params_dict: Dict[str, Union[str, int, float]] = {}
        self.api_key = api_key
        self.kwargs = search_engine_kwargs if search_engine_kwargs else {}
        self.engine = engine
        self.top_k = top_k

    def search(self, query: str, **kwargs) -> List[Document]:
        """
        :param query: Query string.
        :param kwargs: Additional parameters passed to the SerpAPI. For example, you can set 'lr' to 'lang_en'
        to limit the search to English.
        See the [SerpAPI documentation](https://serpapi.com/search-api) for the full list of supported parameters.
        :return: List[Document]
        """
        kwargs = {**self.kwargs, **kwargs}
        top_k = kwargs.pop("top_k", self.top_k)
        url = "https://serpapi.com/search"

        params = {"source": "python", "serp_api_key": self.api_key, "q": query, **kwargs}

        if self.engine:
            params["engine"] = self.engine
        response = requests.get(url, params, timeout=30)

        if response.status_code != 200:
            raise Exception(f"Error while querying {self.__class__.__name__}: {response.text}")

        json_result = json.loads(response.text)
        organic = [
            Document.from_dict(d, field_map={"snippet": "content"})
            for d in json_result["organic_results"]
            if "snippet" in d
        ]
        answer_box = []
        if "answer_box" in json_result:
            answer_dict = json_result["answer_box"]
            for key in ["answer", "snippet_highlighted_words", "snippet", "title"]:
                if key in answer_dict:
                    answer_box_content = answer_dict[key]
                    if isinstance(answer_box_content, list):
                        answer_box_content = answer_box_content[0]
                    answer_box = [
                        Document.from_dict(
                            {
                                "title": answer_dict.get("title", ""),
                                "content": answer_box_content,
                                "link": answer_dict.get("displayed_link", ""),
                            }
                        )
                    ]
                    break

        people_also_search = []
        if "people_also_search_for" in json_result:
            for result in json_result["people_also_search_for"]:
                people_also_search.append(
                    Document.from_dict(
                        {
                            "title": result["title"],
                            "content": result["snippet"] if result.get("snippet") else result["title"],
                            "link": result["link"],
                        }
                    )
                )

        related_questions = []
        if "related_questions" in json_result:
            for result in json_result["related_questions"]:
                related_questions.append(
                    Document.from_dict(
                        {
                            "title": result["title"],
                            "content": result["snippet"] if result.get("snippet") else result["title"],
                            "link": result["link"],
                        }
                    )
                )

        documents = answer_box + organic + people_also_search + related_questions

        logger.debug("SerpAPI returned %s documents for the query '%s'", len(documents), query)
        result_docs = documents[:top_k]
        return self.score_results(result_docs, len(answer_box) > 0)

__init__ 

__init__(api_key: str, top_k: Optional[int] = 10, engine: Optional[str] = 'google', search_engine_kwargs: Optional[Dict[str, Any]] = None)

Parameters:

Name Type Description Default
api_key str

API key for SerpAPI.

 required
top_k Optional[int]

Number of results to return.

 10
engine Optional[str]

Search engine to use, for example google, bing, baidu, duckduckgo, yahoo, yandex. See the SerpAPI documentation for the full list of supported engines.

 'google'
search_engine_kwargs Optional[Dict[str, Any]]

Additional parameters passed to the SerperDev API. For example, you can set 'lr' to 'lang_en' to limit the search to English. See the SerpAPI documentation for the full list of supported parameters.

 None
Source code in pipelines/pipelines/nodes/search_engine/providers.py 
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
def __init__(
    self,
    api_key: str,
    top_k: Optional[int] = 10,
    engine: Optional[str] = "google",
    search_engine_kwargs: Optional[Dict[str, Any]] = None,
):
    """
    :param api_key: API key for SerpAPI.
    :param top_k: Number of results to return.
    :param engine: Search engine to use, for example google, bing, baidu, duckduckgo, yahoo, yandex.
    See the [SerpAPI documentation](https://serpapi.com/search-api) for the full list of supported engines.
    :param search_engine_kwargs: Additional parameters passed to the SerperDev API. For example, you can set 'lr' to 'lang_en'
    to limit the search to English.
    See the [SerpAPI documentation](https://serpapi.com/search-api) for the full list of supported parameters.
    """
    super().__init__()
    self.params_dict: Dict[str, Union[str, int, float]] = {}
    self.api_key = api_key
    self.kwargs = search_engine_kwargs if search_engine_kwargs else {}
    self.engine = engine
    self.top_k = top_k

search 

search(query: str, **kwargs) -> List[Document]

Parameters:

Name Type Description Default
query str

Query string.

 required
kwargs

Additional parameters passed to the SerpAPI. For example, you can set 'lr' to 'lang_en' to limit the search to English. See the SerpAPI documentation for the full list of supported parameters.

 {}

Returns:

Type Description
List[Document]

List[Document]
Source code in pipelines/pipelines/nodes/search_engine/providers.py 
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
def search(self, query: str, **kwargs) -> List[Document]:
    """
    :param query: Query string.
    :param kwargs: Additional parameters passed to the SerpAPI. For example, you can set 'lr' to 'lang_en'
    to limit the search to English.
    See the [SerpAPI documentation](https://serpapi.com/search-api) for the full list of supported parameters.
    :return: List[Document]
    """
    kwargs = {**self.kwargs, **kwargs}
    top_k = kwargs.pop("top_k", self.top_k)
    url = "https://serpapi.com/search"

    params = {"source": "python", "serp_api_key": self.api_key, "q": query, **kwargs}

    if self.engine:
        params["engine"] = self.engine
    response = requests.get(url, params, timeout=30)

    if response.status_code != 200:
        raise Exception(f"Error while querying {self.__class__.__name__}: {response.text}")

    json_result = json.loads(response.text)
    organic = [
        Document.from_dict(d, field_map={"snippet": "content"})
        for d in json_result["organic_results"]
        if "snippet" in d
    ]
    answer_box = []
    if "answer_box" in json_result:
        answer_dict = json_result["answer_box"]
        for key in ["answer", "snippet_highlighted_words", "snippet", "title"]:
            if key in answer_dict:
                answer_box_content = answer_dict[key]
                if isinstance(answer_box_content, list):
                    answer_box_content = answer_box_content[0]
                answer_box = [
                    Document.from_dict(
                        {
                            "title": answer_dict.get("title", ""),
                            "content": answer_box_content,
                            "link": answer_dict.get("displayed_link", ""),
                        }
                    )
                ]
                break

    people_also_search = []
    if "people_also_search_for" in json_result:
        for result in json_result["people_also_search_for"]:
            people_also_search.append(
                Document.from_dict(
                    {
                        "title": result["title"],
                        "content": result["snippet"] if result.get("snippet") else result["title"],
                        "link": result["link"],
                    }
                )
            )

    related_questions = []
    if "related_questions" in json_result:
        for result in json_result["related_questions"]:
            related_questions.append(
                Document.from_dict(
                    {
                        "title": result["title"],
                        "content": result["snippet"] if result.get("snippet") else result["title"],
                        "link": result["link"],
                    }
                )
            )

    documents = answer_box + organic + people_also_search + related_questions

    logger.debug("SerpAPI returned %s documents for the query '%s'", len(documents), query)
    result_docs = documents[:top_k]
    return self.score_results(result_docs, len(answer_box) > 0)

SerperDev

Serper.dev is a search engine that provides a REST API to access search results from Google. See the Serper.dev website for more details.

Source code in pipelines/pipelines/nodes/search_engine/providers.py 
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
class SerperDev(SearchEngine):
    """
    Serper.dev is a search engine that provides a REST API to access search results from Google. See the [Serper.dev website](https://serper.dev.com/) for more details.
    """

    def __init__(
        self,
        api_key: str,
        top_k: Optional[int] = 10,
        engine: Optional[str] = "google",
        search_engine_kwargs: Optional[Dict[str, Any]] = None,
    ):
        """
        :param api_key: API key for Serper.dev API.
        :param top_k: Number of results to return.
        :param engine: Search engine to use, only supports Google.
        :param search_engine_kwargs: Additional parameters passed to the SerperDev API. For example, you can set 'hl' to 'en'
        to set the search results language to English.
        See the [Serper.dev documentation](https://serper.dev/playground) for the full list of supported parameters.
        """
        super().__init__()
        self.params_dict: Dict[str, Union[str, int, float]] = {}
        self.api_key = api_key
        self.kwargs = search_engine_kwargs if search_engine_kwargs else {}
        self.engine = engine
        self.top_k = top_k

    def search(self, query: str, **kwargs) -> List[Document]:
        """
        :param query: Query string.
        :param kwargs: Additional parameters passed to the Serper.dev API. For example, you can set 'hl' to 'en'
        to set the search results language to English.
        See the [Serper.dev documentation](https://serper.dev/playground) for the full list of supported parameters.
        :return: List[Document]
        """
        kwargs = {**self.kwargs, **kwargs}
        top_k = kwargs.pop("top_k", self.top_k)
        url = "https://google.serper.dev/search"

        params = {"q": query, **kwargs}

        headers = {"X-API-KEY": self.api_key, "Content-Type": "application/json"}

        response = requests.post(url, headers=headers, json=params, timeout=30)

        if response.status_code != 200:
            raise Exception(f"Error while querying {self.__class__.__name__}: {response.text}")

        json_result = json.loads(response.text)
        organic = [
            Document.from_dict(d, field_map={"snippet": "content"}) for d in json_result["organic"] if "snippet" in d
        ]
        answer_box = []
        if "answerBox" in json_result:
            answer_dict = json_result["answerBox"]
            for key in ["answer", "snippetHighlighted", "snippet", "title"]:
                if key in answer_dict:
                    answer_box_content = answer_dict[key]
                    if isinstance(answer_box_content, list):
                        answer_box_content = answer_box_content[0]
                    answer_box = [
                        Document.from_dict(
                            {
                                "title": answer_dict.get("title", ""),
                                "content": answer_box_content,
                                "link": answer_dict.get("link", ""),
                            }
                        )
                    ]
                    break

        people_also_search = []
        if "peopleAlsoSearchFor" in json_result:
            for result in json_result["peopleAlsoSearchFor"]:
                people_also_search.append(
                    Document.from_dict(
                        {
                            "title": result["title"],
                            "content": result["snippet"] if result.get("snippet") else result["title"],
                            "link": result["link"],
                        }
                    )
                )

        related_searches = []
        if "relatedSearches" in json_result:
            for result in json_result["relatedSearches"]:
                related_searches.append(Document.from_dict({"content": result.get("query", "")}))

        related_questions = []
        if "peopleAlsoAsk" in json_result:
            for result in json_result["peopleAlsoAsk"]:
                related_questions.append(
                    Document.from_dict(
                        {
                            "title": result["title"],
                            "content": result["snippet"] if result.get("snippet") else result["title"],
                            "link": result["link"],
                        }
                    )
                )

        documents = answer_box + organic + people_also_search + related_searches + related_questions

        logger.debug("Serper.dev API returned %s documents for the query '%s'", len(documents), query)
        result_docs = documents[:top_k]
        return self.score_results(result_docs, len(answer_box) > 0)

__init__ 

__init__(api_key: str, top_k: Optional[int] = 10, engine: Optional[str] = 'google', search_engine_kwargs: Optional[Dict[str, Any]] = None)

Parameters:

Name Type Description Default
api_key str

API key for Serper.dev API.

 required
top_k Optional[int]

Number of results to return.

 10
engine Optional[str]

Search engine to use, only supports Google.

 'google'
search_engine_kwargs Optional[Dict[str, Any]]

Additional parameters passed to the SerperDev API. For example, you can set 'hl' to 'en' to set the search results language to English. See the Serper.dev documentation for the full list of supported parameters.

 None
Source code in pipelines/pipelines/nodes/search_engine/providers.py 
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
def __init__(
    self,
    api_key: str,
    top_k: Optional[int] = 10,
    engine: Optional[str] = "google",
    search_engine_kwargs: Optional[Dict[str, Any]] = None,
):
    """
    :param api_key: API key for Serper.dev API.
    :param top_k: Number of results to return.
    :param engine: Search engine to use, only supports Google.
    :param search_engine_kwargs: Additional parameters passed to the SerperDev API. For example, you can set 'hl' to 'en'
    to set the search results language to English.
    See the [Serper.dev documentation](https://serper.dev/playground) for the full list of supported parameters.
    """
    super().__init__()
    self.params_dict: Dict[str, Union[str, int, float]] = {}
    self.api_key = api_key
    self.kwargs = search_engine_kwargs if search_engine_kwargs else {}
    self.engine = engine
    self.top_k = top_k

search 

search(query: str, **kwargs) -> List[Document]

Parameters:

Name Type Description Default
query str

Query string.

 required
kwargs

Additional parameters passed to the Serper.dev API. For example, you can set 'hl' to 'en' to set the search results language to English. See the Serper.dev documentation for the full list of supported parameters.

 {}

Returns:

Type Description
List[Document]

List[Document]
Source code in pipelines/pipelines/nodes/search_engine/providers.py 
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
def search(self, query: str, **kwargs) -> List[Document]:
    """
    :param query: Query string.
    :param kwargs: Additional parameters passed to the Serper.dev API. For example, you can set 'hl' to 'en'
    to set the search results language to English.
    See the [Serper.dev documentation](https://serper.dev/playground) for the full list of supported parameters.
    :return: List[Document]
    """
    kwargs = {**self.kwargs, **kwargs}
    top_k = kwargs.pop("top_k", self.top_k)
    url = "https://google.serper.dev/search"

    params = {"q": query, **kwargs}

    headers = {"X-API-KEY": self.api_key, "Content-Type": "application/json"}

    response = requests.post(url, headers=headers, json=params, timeout=30)

    if response.status_code != 200:
        raise Exception(f"Error while querying {self.__class__.__name__}: {response.text}")

    json_result = json.loads(response.text)
    organic = [
        Document.from_dict(d, field_map={"snippet": "content"}) for d in json_result["organic"] if "snippet" in d
    ]
    answer_box = []
    if "answerBox" in json_result:
        answer_dict = json_result["answerBox"]
        for key in ["answer", "snippetHighlighted", "snippet", "title"]:
            if key in answer_dict:
                answer_box_content = answer_dict[key]
                if isinstance(answer_box_content, list):
                    answer_box_content = answer_box_content[0]
                answer_box = [
                    Document.from_dict(
                        {
                            "title": answer_dict.get("title", ""),
                            "content": answer_box_content,
                            "link": answer_dict.get("link", ""),
                        }
                    )
                ]
                break

    people_also_search = []
    if "peopleAlsoSearchFor" in json_result:
        for result in json_result["peopleAlsoSearchFor"]:
            people_also_search.append(
                Document.from_dict(
                    {
                        "title": result["title"],
                        "content": result["snippet"] if result.get("snippet") else result["title"],
                        "link": result["link"],
                    }
                )
            )

    related_searches = []
    if "relatedSearches" in json_result:
        for result in json_result["relatedSearches"]:
            related_searches.append(Document.from_dict({"content": result.get("query", "")}))

    related_questions = []
    if "peopleAlsoAsk" in json_result:
        for result in json_result["peopleAlsoAsk"]:
            related_questions.append(
                Document.from_dict(
                    {
                        "title": result["title"],
                        "content": result["snippet"] if result.get("snippet") else result["title"],
                        "link": result["link"],
                    }
                )
            )

    documents = answer_box + organic + people_also_search + related_searches + related_questions

    logger.debug("Serper.dev API returned %s documents for the query '%s'", len(documents), query)
    result_docs = documents[:top_k]
    return self.score_results(result_docs, len(answer_box) > 0)

pipelines.pipelines.nodes.search_engine.web

WebSearch

WebSearch queries a search engine and retrieves results as a list of Documents. WebSearch abstracts away the details of the underlying search engine provider, provides common interface for all providers, and makes it possible to use various search engines.

WebSerach currently supports the following search engines providers (bridges): - SerperDev (default) - SearchApi - SerpAPI - BingAPI

Source code in pipelines/pipelines/nodes/search_engine/web.py 
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
class WebSearch(BaseComponent):
    """
    WebSearch queries a search engine and retrieves results as a list of Documents. WebSearch abstracts away the details
    of the underlying search engine provider, provides common interface for all providers, and makes it possible to use various
    search engines.

    WebSerach currently supports the following search engines providers (bridges):
    - SerperDev (default)
    - SearchApi
    - SerpAPI
    - BingAPI

    """

    outgoing_edges = 1

    def __init__(
        self,
        api_key: str,
        top_k: Optional[int] = 10,
        search_engine_provider: Union[str, SearchEngine] = "SerpAPI",
        engine: Optional[str] = "google",
        search_engine_kwargs: Optional[Dict[str, Any]] = None,
    ):
        """
        :param api_key: API key for the search engine provider.
        :param search_engine_provider: Name of the search engine provider class, see `providers.py` for a list of
        supported providers.
        :param search_engine_kwargs: Additional parameters to pass to the search engine provider.
        """
        super().__init__()
        if isinstance(search_engine_provider, str):
            # try to find the provider class
            search_path = [f"pipelines.nodes.search_engine.providers.{search_engine_provider}", search_engine_provider]
            klass: Type[SearchEngine] = next((pydoc.locate(path) for path in search_path), None)  # type: ignore

            if not klass:
                raise ValueError(
                    f"Could not locate the SearchEngine class with the name {search_engine_provider}. "
                    f"Make sure you pass the full path to the class."
                )
            if not issubclass(klass, SearchEngine):
                raise ValueError(f"Class {search_engine_provider} is not a subclass of SearchEngine.")
            self.search_engine = klass(api_key=api_key, top_k=top_k, engine=engine, search_engine_kwargs=search_engine_kwargs)  # type: ignore
        elif isinstance(search_engine_provider, SearchEngine):
            self.search_engine = search_engine_provider
        else:
            raise ValueError(
                "search_engine_provider must be either a string (SearchEngine class name) or a SearchEngine instance."
            )

    def run(
        self,
        query: Optional[str] = None,
        file_paths: Optional[List[str]] = None,
        labels: Optional[MultiLabel] = None,
        documents: Optional[List[Document]] = None,
        meta: Optional[dict] = None,
    ) -> Tuple[Dict, str]:
        """
        Search the search engine for the given query and return the results. Only the query parameter is used.
        :param query: The query to search for.

        :return: List of search results as documents.
        """
        # query is a required parameter for search, we need to keep the signature of run() the same as in other nodes
        if not query:
            raise ValueError("WebSearch run requires the `query` parameter")
        return {"documents": self.search_engine.search(query)}, "output_1"

    def run_batch(
        self,
        queries: Optional[Union[str, List[str]]] = None,
        file_paths: Optional[List[str]] = None,
        labels: Optional[Union[MultiLabel, List[MultiLabel]]] = None,
        documents: Optional[Union[List[Document], List[List[Document]]]] = None,
        meta: Optional[Union[Dict[str, Any], List[Dict[str, Any]]]] = None,
        params: Optional[dict] = None,
        debug: Optional[bool] = None,
    ):
        results = []
        if isinstance(queries, str):
            queries = [queries]
        elif not isinstance(queries, list):
            raise ValueError("WebSearch run_batch requires the `queries` parameter to be Union[str, List[str]]")
        for query in queries:
            results.append(self.search_engine.search(query))
        return {"documents": results}, "output_1"

__init__ 

__init__(api_key: str, top_k: Optional[int] = 10, search_engine_provider: Union[str, SearchEngine] = 'SerpAPI', engine: Optional[str] = 'google', search_engine_kwargs: Optional[Dict[str, Any]] = None)

Parameters:

Name Type Description Default
api_key str

API key for the search engine provider.

 required
search_engine_provider Union[str, SearchEngine]

Name of the search engine provider class, see providers.py for a list of supported providers.

 'SerpAPI'
search_engine_kwargs Optional[Dict[str, Any]]

Additional parameters to pass to the search engine provider.

 None
Source code in pipelines/pipelines/nodes/search_engine/web.py 
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
def __init__(
    self,
    api_key: str,
    top_k: Optional[int] = 10,
    search_engine_provider: Union[str, SearchEngine] = "SerpAPI",
    engine: Optional[str] = "google",
    search_engine_kwargs: Optional[Dict[str, Any]] = None,
):
    """
    :param api_key: API key for the search engine provider.
    :param search_engine_provider: Name of the search engine provider class, see `providers.py` for a list of
    supported providers.
    :param search_engine_kwargs: Additional parameters to pass to the search engine provider.
    """
    super().__init__()
    if isinstance(search_engine_provider, str):
        # try to find the provider class
        search_path = [f"pipelines.nodes.search_engine.providers.{search_engine_provider}", search_engine_provider]
        klass: Type[SearchEngine] = next((pydoc.locate(path) for path in search_path), None)  # type: ignore

        if not klass:
            raise ValueError(
                f"Could not locate the SearchEngine class with the name {search_engine_provider}. "
                f"Make sure you pass the full path to the class."
            )
        if not issubclass(klass, SearchEngine):
            raise ValueError(f"Class {search_engine_provider} is not a subclass of SearchEngine.")
        self.search_engine = klass(api_key=api_key, top_k=top_k, engine=engine, search_engine_kwargs=search_engine_kwargs)  # type: ignore
    elif isinstance(search_engine_provider, SearchEngine):
        self.search_engine = search_engine_provider
    else:
        raise ValueError(
            "search_engine_provider must be either a string (SearchEngine class name) or a SearchEngine instance."
        )

run 

run(query: Optional[str] = None, file_paths: Optional[List[str]] = None, labels: Optional[MultiLabel] = None, documents: Optional[List[Document]] = None, meta: Optional[dict] = None) -> Tuple[Dict, str]

Search the search engine for the given query and return the results. Only the query parameter is used.

Parameters:

Name Type Description Default
query Optional[str]

The query to search for.

 None

Returns:

Type Description
Tuple[Dict, str]

List of search results as documents.
Source code in pipelines/pipelines/nodes/search_engine/web.py 
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
def run(
    self,
    query: Optional[str] = None,
    file_paths: Optional[List[str]] = None,
    labels: Optional[MultiLabel] = None,
    documents: Optional[List[Document]] = None,
    meta: Optional[dict] = None,
) -> Tuple[Dict, str]:
    """
    Search the search engine for the given query and return the results. Only the query parameter is used.
    :param query: The query to search for.

    :return: List of search results as documents.
    """
    # query is a required parameter for search, we need to keep the signature of run() the same as in other nodes
    if not query:
        raise ValueError("WebSearch run requires the `query` parameter")
    return {"documents": self.search_engine.search(query)}, "output_1"