Migrate http client from requests to httpx async client

juriscraper/AbstractSite.py

@@ -3,7 +3,7 @@
 from datetime import date, datetime
 
 import certifi
-import requests
+import httpx
 
 from juriscraper.lib.date_utils import fix_future_year_typo, json_date_handler
 from juriscraper.lib.exceptions import InsanityException
@@ -33,7 +33,7 @@ class AbstractSite:
     Should not contain lists that can't be sorted by the _date_sort function.
     """
 
-    def __init__(self, cnt=None):
+    def __init__(self, cnt=None, user_agent="Juriscraper", **kwargs):
         super().__init__()
 
         # Computed metadata
@@ -44,10 +44,12 @@ def __init__(self, cnt=None):
         self.downloader_executed = False
         self.cookies = {}
         self.cnt = cnt or CaseNameTweaker()
+        self.user_agent = user_agent
+        kwargs.setdefault("http2", True)
         self.request = {
             "verify": certifi.where(),
-            "session": requests.session(),
-            "headers": {"User-Agent": "Juriscraper"},
+            "session": httpx.AsyncClient(**kwargs),
+            "headers": {"User-Agent": self.user_agent},
             # Disable CDN caching on sites like SCOTUS (ahem)
             "cache-control": "no-cache, no-store, max-age=1",
             "parameters": {},
@@ -65,8 +67,8 @@ def __init__(self, cnt=None):
         self._req_attrs = []
         self._all_attrs = []
 
-    def __del__(self):
-        self.close_session()
+    async def __aexit__(self):
+        await self.close_session()
 
     def __str__(self):
         out = []
@@ -84,9 +86,9 @@ def __getitem__(self, i):
     def __len__(self):
         return len(self.case_names)
 
-    def close_session(self):
+    async def close_session(self):
         if self.request["session"]:
-            self.request["session"].close()
+            await self.request["session"].aclose()
 
     def _make_item(self, i):
         """Using i, convert a single item into a dict. This is effectively a
@@ -344,23 +346,23 @@ def _process_request_parameters(self, parameters={}):
             del parameters["verify"]
         self.request["parameters"] = parameters
 
-    def _request_url_get(self, url):
+    async def _request_url_get(self, url):
         """Execute GET request and assign appropriate request dictionary
         values
         """
         self.request["url"] = url
-        self.request["response"] = self.request["session"].get(
+        self.request["response"] = await self.request["session"].get(
             url,
             headers=self.request["headers"],
             verify=self.request["verify"],
             timeout=60,
             **self.request["parameters"],
         )
 
-    def _request_url_post(self, url):
+    async def _request_url_post(self, url):
         """Execute POST request and assign appropriate request dictionary values"""
         self.request["url"] = url
-        self.request["response"] = self.request["session"].post(
+        self.request["response"] = await self.request["session"].post(
             url,
             headers=self.request["headers"],
             verify=self.request["verify"],

juriscraper/DeferringList.py

@@ -1,3 +1,6 @@
+import asyncio
+import inspect
+
 from juriscraper.AbstractSite import logger
 
 
@@ -42,7 +45,13 @@ def __getitem__(self, item):
             logger.info(
                 f"Getting deferred value from seed: {self._data[item]}"
             )
-            new_val = self._fetching_function(self._data[item])
+            if inspect.isawaitable(self._fetching_function):
+                loop = asyncio.get_event_loop()
+                new_val = loop.run_until_complete(
+                    self._fetching_function(self._data[item])
+                )
+            else:
+                new_val = self._fetching_function(self._data[item])
             self._data[item] = new_val
             self._fetched_items[item] = True
             return new_val

juriscraper/OpinionSiteLinearWebDriven.py

@@ -7,6 +7,8 @@ def __init__(self, *args, **kwargs):
         super().__init__(*args, **kwargs)
         WebDriven.__init__(self, args, kwargs)
 
+    async def __aexit__(self):
+        await self.close_session()
+
     def __del__(self):
-        self.close_session()
         self.close_webdriver_session()

juriscraper/OpinionSiteWebDriven.py

@@ -7,6 +7,8 @@ def __init__(self, *args, **kwargs):
         super().__init__(*args, **kwargs)
         WebDriven.__init__(self, args, kwargs)
 
+    async def __aexit__(self):
+        await self.close_session()
+
     def __del__(self):
-        self.close_session()
         self.close_webdriver_session()

juriscraper/OralArgumentSiteLinearWebDriven.py

@@ -7,6 +7,8 @@ def __init__(self, *args, **kwargs):
         super().__init__(*args, **kwargs)
         WebDriven.__init__(self, args, kwargs)
 
+    async def __aexit__(self):
+        await self.close_session()
+
     def __del__(self):
-        self.close_session()
         self.close_webdriver_session()

juriscraper/fdsys/FDSysSite.py

@@ -5,19 +5,20 @@
 from datetime import date
 from pprint import pprint
 
-import requests
+import httpx
+from httpx import InvalidURL
 from lxml import etree
-from requests.exceptions import MissingSchema
 
 from juriscraper.AbstractSite import AbstractSite
 
 
-def get_tree(url):
+async def get_tree(url, **kwargs):
     try:
-        response = requests.get(url, stream=True)
-        response.raw.decode_content = True
-        return etree.parse(response.raw)
-    except MissingSchema:
+        kwargs.setdefault("http2", True)
+        async with httpx.AsyncClient(**kwargs) as client:
+            response = await client.get(url)
+            return etree.parse(await response.aread())
+    except InvalidURL:
         return etree.parse(url)
 
 
@@ -160,23 +161,25 @@ def __getitem__(self, i):
     def __len__(self):
         return len(xpath(self.html, "//s:loc/text()"))
 
-    def save_mods_file(self, url):
+    async def save_mods_file(self, url, **kwargs):
         mods_url = FDSysModsContent._get_mods_file_url(url)
         name = "-".join(mods_url.split("/")[-2].split("-")[1:])
-        with open(f"./examples/2006/{name}.xml", "w") as handle:
-            response = requests.get(mods_url, stream=True)
-            for block in response.iter_content(1024):
-                handle.write(block)
-
-    def _download(self, request_dict={}):
+        with open(f"./examples/2006/{name}.xml", "wb") as handle:
+            kwargs.setdefault("http2", True)
+            async with httpx.AsyncClient(**kwargs) as client:
+                async with client.stream("GET", mods_url) as response:
+                    async for block in response.aiter_bytes():
+                        handle.write(block)
+
+    async def _download(self, request_dict={}):
         """
         it actually builds an XML tree
         """
-        return get_tree(self.url)
+        return await get_tree(self.url)
 
-    def _download_backwards(self, year):
+    async def _download_backwards(self, year):
         self.url = self.base_url.format(year=year)
-        self.html = self._download()
+        self.html = await self._download()
         if self.html is not None:
             # Setting status is important because it prevents the download
             # function from being run a second time by the parse method.
@@ -185,10 +188,10 @@ def _download_backwards(self, year):
     def _check_sanity(self):
         pass
 
-    def parse(self):
+    async def parse(self):
         if self.status is None:
             # Run the downloader if it hasn't been run already
-            self.html = self._download()
+            self.html = await self._download()
         return self
 
 

juriscraper/fdsys/scrape_court_names.py

@@ -1,12 +1,17 @@
+import asyncio
 import json
 from pprint import pprint
 
-import requests
+import httpx
 from lxml import etree, html
 
 
-def get_court_names():
-    response = requests.get("https://www.courtlistener.com/api/jurisdictions/")
+async def get_court_names(**kwargs):
+    kwargs.setdefault("http2", True)
+    async with httpx.AsyncClient(**kwargs) as client:
+        response = await client.get(
+            "https://www.courtlistener.com/api/jurisdictions/"
+        )
     tree = html.fromstring(response.text)
 
     data = dict()
@@ -21,13 +26,14 @@ def get_court_names():
         json.dump(data, f)
 
 
-def get_fdsys_court_names():
-    response = requests.get(
-        "https://www.gpo.gov/smap/fdsys/sitemap_2014/2014_USCOURTS_sitemap.xml",
-        stream=True,
-    )
-    response.raw.decode_content = True
-    tree = etree.parse(response.raw)
+async def get_fdsys_court_names(**kwargs):
+    kwargs.setdefault("http2", True)
+    async with httpx.AsyncClient(**kwargs) as client:
+        response = await client.get(
+            "https://www.gpo.gov/smap/fdsys/sitemap_2014/2014_USCOURTS_sitemap.xml"
+        )
+        tree = etree.parse(await response.aread())
+
     data = dict()
 
     for url in tree.xpath(
@@ -47,4 +53,4 @@ def get_fdsys_court_names():
 
 if __name__ == "__main__":
     # get_court_names()
-    get_fdsys_court_names()
+    asyncio.run(get_fdsys_court_names())

juriscraper/lasc/http.py

@@ -1,22 +1,20 @@
-import requests
+import httpx
 from lxml.html import fromstring
 
 from ..lib.log_tools import make_default_logger
 
-requests.packages.urllib3.disable_warnings(
-    requests.packages.urllib3.exceptions.InsecureRequestWarning
-)
-
 logger = make_default_logger()
 
 
-class LASCSession(requests.Session):
+class LASCSession(httpx.AsyncClient):
     """
-    A requests.Session object with special tooling to handle the Los Angeles
+    A httpx.AsyncClient object with special tooling to handle the Los Angeles
     Superior Court Media Access portal.
     """
 
-    def __init__(self, username=None, password=None):
+    def __init__(
+        self, username=None, password=None, user_agent="Juriscraper", **kwargs
+    ):
         """
         Instantiate a new LASC HTTP Session with some Juriscraper defaults.
         This method requires credentials from the media access portal.
@@ -25,7 +23,7 @@ def __init__(self, username=None, password=None):
         :param password: MAP password
         :return: A LASCSession object
         """
-        super().__init__()
+        super().__init__(**kwargs)
 
         self.html = None
 
@@ -53,34 +51,35 @@ def __init__(self, username=None, password=None):
             "password": password,
             "request_type": "RESPONSE",
         }
+        self.user_agent = user_agent
         self.headers = {
             "Origin": ms_base_url,
-            "User-Agent": "Juriscraper",
+            "User-Agent": self.user_agent,
         }
 
-    def get(self, url, auto_login=False, **kwargs):
-        """Overrides request.Session.get with session retry logic.
+    async def get(self, url, auto_login=False, **kwargs):
+        """Overrides httpx.AsyncClient.get with session retry logic.
 
         :param url: url string to GET
         :param auto_login: Whether the auto-login procedure should happen.
-        :return: requests.Response
+        :return: httpx.Response
         """
         kwargs.setdefault("timeout", 30)
         kwargs.setdefault("params", {"p": "B2C_1_Media-LASC-SUSI"})
 
-        return super().get(url, **kwargs)
+        return await super().get(url, **kwargs)
 
-    def post(self, url, auto_login=False, **kwargs):
-        """Overrides request.Session.post with session retry logic.
+    async def post(self, url, auto_login=False, **kwargs):
+        """Overrides httpx.AsyncClient.post with session retry logic.
 
         :param url: url string to GET
         :param auto_login: Whether the auto-login procedure should happen.
-        :return: requests.Response
+        :return: httpx.Response
         """
         kwargs.setdefault("timeout", 30)
         kwargs.setdefault("params", {"p": "B2C_1_Media-LASC-SUSI"})
 
-        return super().post(url, **kwargs)
+        return await super().post(url, **kwargs)
 
     @staticmethod
     def _parse_new_html_for_keys(r):
@@ -89,7 +88,7 @@ def _parse_new_html_for_keys(r):
         This method parses the HTML after the first login page and identifies
         the parameter values required for the next step.
 
-        :param r: A request.Response object
+        :param r: A httpx.Response object
         :return: A dict containing the needed keys
         """
         html = fromstring(r.text)
@@ -103,7 +102,7 @@ def _parse_new_html_for_keys(r):
     def _check_login(r):
         """Check that the login succeeded
 
-        :param r: A request.Response object
+        :param r: A httpx.Response object
         :return: None
         :raises LASCLoginException
         """
@@ -121,7 +120,7 @@ def _check_login(r):
     def _update_header_token(self, r):
         self.headers["X-CSRF-TOKEN"] = r.text.split("csrf")[1].split('"')[2]
 
-    def login(self):
+    async def login(self):
         """Log into the LASC Media Access Portal
         The process is tricky, requiring two GET requests, each of which
         returns HTML or JSON that is parsed for values to send in a subsequent
@@ -326,20 +325,20 @@ def login(self):
         """
 
         logger.info("Logging into MAP has begun")
-        r = self.get(self.login_url)
+        r = await self.get(self.login_url)
         self._update_header_token(r)
 
         # Call part one of Microsoft login API
-        r = self.post(self.api_url1, data=self.login_data)
+        r = await self.post(self.api_url1, data=self.login_data)
         self._check_login(r)
 
         # Call part two of Microsoft login API - Redirect
-        r = self.get(self.api_url2)
+        r = await self.get(self.api_url2)
 
         # Finalize login with post into LA MAP site
         parsed_keys = self._parse_new_html_for_keys(r)
 
-        self.post(self.signin_url, data=parsed_keys)
+        await self.post(self.signin_url, data=parsed_keys)
 
         logger.info("Successfully Logged into MAP")