262 lines
9.6 KiB
Python
262 lines
9.6 KiB
Python
"""
|
|
This module interfaces the Wayback Machine's CDX server API.
|
|
|
|
The module has WaybackMachineCDXServerAPI which should be used by the users of
|
|
this module to consume the CDX server API.
|
|
|
|
WaybackMachineCDXServerAPI has a snapshot method that yields the snapshots, and
|
|
the snapshots are yielded as instances of the CDXSnapshot class.
|
|
"""
|
|
|
|
|
|
from typing import Dict, Generator, List, Optional, cast
|
|
|
|
from .cdx_snapshot import CDXSnapshot
|
|
from .cdx_utils import (
|
|
check_collapses,
|
|
check_filters,
|
|
check_match_type,
|
|
full_url,
|
|
get_response,
|
|
get_total_pages,
|
|
)
|
|
from .exceptions import WaybackError
|
|
from .utils import DEFAULT_USER_AGENT
|
|
|
|
|
|
class WaybackMachineCDXServerAPI:
|
|
"""
|
|
Class that interfaces the CDX server API of the Wayback Machine.
|
|
|
|
snapshot() returns a generator that can be iterated upon by the end-user,
|
|
the generator returns the snapshots/entries as instance of CDXSnapshot to
|
|
make the usage easy, just use '.' to get any attribute as the attributes are
|
|
accessible via a dot ".".
|
|
"""
|
|
|
|
# start_timestamp: from, can not use from as it's a keyword
|
|
# end_timestamp: to, not using to as can not use from
|
|
def __init__(
|
|
self,
|
|
url: str,
|
|
user_agent: str = DEFAULT_USER_AGENT,
|
|
start_timestamp: Optional[str] = None,
|
|
end_timestamp: Optional[str] = None,
|
|
filters: Optional[List[str]] = None,
|
|
match_type: Optional[str] = None,
|
|
gzip: Optional[str] = None,
|
|
collapses: Optional[List[str]] = None,
|
|
limit: Optional[str] = None,
|
|
max_tries: int = 3,
|
|
) -> None:
|
|
self.url = str(url).strip().replace(" ", "%20")
|
|
self.user_agent = user_agent
|
|
self.start_timestamp = None if start_timestamp is None else str(start_timestamp)
|
|
self.end_timestamp = None if end_timestamp is None else str(end_timestamp)
|
|
self.filters = [] if filters is None else filters
|
|
check_filters(self.filters)
|
|
self.match_type = None if match_type is None else str(match_type).strip()
|
|
check_match_type(self.match_type, self.url)
|
|
self.gzip = gzip
|
|
self.collapses = [] if collapses is None else collapses
|
|
check_collapses(self.collapses)
|
|
self.limit = 25000 if limit is None else limit
|
|
self.max_tries = max_tries
|
|
self.last_api_request_url: Optional[str] = None
|
|
self.use_page = False
|
|
self.endpoint = "https://web.archive.org/cdx/search/cdx"
|
|
|
|
def cdx_api_manager(
|
|
self, payload: Dict[str, str], headers: Dict[str, str], use_page: bool = False
|
|
) -> Generator[str, None, None]:
|
|
"""
|
|
Manages the API calls for the instance, it automatically selects the best
|
|
parameters by looking as the query of the end-user. For bigger queries
|
|
automatically use the CDX pagination API and for smaller queries use the
|
|
normal API.
|
|
|
|
CDX Server API is a complex API and to make it easy for the end user to
|
|
consume it the CDX manager(this method) handles the selection of the
|
|
API output, whether to use the pagination API or not.
|
|
|
|
For doing large/bulk queries, the use of the Pagination API is
|
|
recommended by the Wayback Machine authors. And it determines if the
|
|
query would be large or not by using the showNumPages=true parameter,
|
|
this tells the number of pages of CDX DATA that the pagination API
|
|
will return.
|
|
|
|
If the number of page is less than 2 we use the normal non-pagination
|
|
API as the pagination API is known to lag and for big queries it should
|
|
not matter but for queries where the number of pages are less this
|
|
method chooses accuracy over the pagination API.
|
|
"""
|
|
# number of pages that will returned by the pagination API.
|
|
# get_total_pages adds the showNumPages=true param to pagination API
|
|
# requests.
|
|
# This is a special query that will return a single number indicating
|
|
# the number of pages.
|
|
total_pages = get_total_pages(self.url, self.user_agent)
|
|
|
|
if use_page is True and total_pages >= 2:
|
|
blank_pages = 0
|
|
for i in range(total_pages):
|
|
payload["page"] = str(i)
|
|
|
|
url = full_url(self.endpoint, params=payload)
|
|
res = get_response(url, headers=headers)
|
|
if isinstance(res, Exception):
|
|
raise res
|
|
|
|
self.last_api_request_url = url
|
|
text = res.text
|
|
if len(text) == 0:
|
|
blank_pages += 1
|
|
|
|
if blank_pages >= 2:
|
|
break
|
|
|
|
yield text
|
|
else:
|
|
payload["showResumeKey"] = "true"
|
|
payload["limit"] = str(self.limit)
|
|
resume_key = None
|
|
more = True
|
|
while more:
|
|
if resume_key:
|
|
payload["resumeKey"] = resume_key
|
|
|
|
url = full_url(self.endpoint, params=payload)
|
|
res = get_response(url, headers=headers)
|
|
if isinstance(res, Exception):
|
|
raise res
|
|
|
|
self.last_api_request_url = url
|
|
|
|
text = res.text.strip()
|
|
lines = text.splitlines()
|
|
|
|
more = False
|
|
|
|
if len(lines) >= 3:
|
|
|
|
second_last_line = lines[-2]
|
|
|
|
if len(second_last_line) == 0:
|
|
|
|
resume_key = lines[-1].strip()
|
|
text = text.replace(resume_key, "", 1).strip()
|
|
more = True
|
|
|
|
yield text
|
|
|
|
def add_payload(self, payload: Dict[str, str]) -> None:
|
|
"""
|
|
Adds the payload to the payload dictionary.
|
|
"""
|
|
if self.start_timestamp:
|
|
payload["from"] = self.start_timestamp
|
|
|
|
if self.end_timestamp:
|
|
payload["to"] = self.end_timestamp
|
|
|
|
if self.gzip is None:
|
|
payload["gzip"] = "false"
|
|
|
|
if self.match_type:
|
|
payload["matchType"] = self.match_type
|
|
|
|
if self.filters and len(self.filters) > 0:
|
|
for i, _filter in enumerate(self.filters):
|
|
payload["filter" + str(i)] = _filter
|
|
|
|
if self.collapses and len(self.collapses) > 0:
|
|
for i, collapse in enumerate(self.collapses):
|
|
payload["collapse" + str(i)] = collapse
|
|
|
|
payload["url"] = self.url
|
|
|
|
def snapshots(self) -> Generator[CDXSnapshot, None, None]:
|
|
"""
|
|
This function yields the CDX data lines as snapshots.
|
|
|
|
As it is a generator it exhaustible, the reason that this is
|
|
a generator and not a list are:
|
|
|
|
a) CDX server API can return millions of entries for a query and list
|
|
is not suitable for such cases.
|
|
|
|
b) Preventing memory usage issues, as told before this method may yield
|
|
millions of records for some queries and your system may not have enough
|
|
memory for such a big list. Also Remember this if outputing to Jupyter
|
|
Notebooks.
|
|
|
|
The objects yielded by this method are instance of CDXSnapshot class,
|
|
you can access the attributes of the entries as the attribute of the instance
|
|
itself.
|
|
"""
|
|
payload: Dict[str, str] = {}
|
|
headers = {"User-Agent": self.user_agent}
|
|
|
|
self.add_payload(payload)
|
|
|
|
if not self.start_timestamp or self.end_timestamp:
|
|
self.use_page = True
|
|
|
|
if self.collapses != []:
|
|
self.use_page = False
|
|
|
|
entries = self.cdx_api_manager(payload, headers, use_page=self.use_page)
|
|
|
|
for entry in entries:
|
|
|
|
if entry.isspace() or len(entry) <= 1 or not entry:
|
|
continue
|
|
|
|
# each line is a snapshot aka entry of the CDX server API.
|
|
# We are able to split the page by lines because it only
|
|
# splits the lines on a sinlge page and not all the entries
|
|
# at once, thus there should be no issues of too much memory usage.
|
|
snapshot_list = entry.split("\n")
|
|
|
|
for snapshot in snapshot_list:
|
|
|
|
# 14 + 32 == 46 ( timestamp + digest ), ignore the invalid entries.
|
|
# they are invalid if their length is smaller than sum of length
|
|
# of a standard wayback_timestamp and standard digest of an entry.
|
|
if len(snapshot) < 46:
|
|
continue
|
|
|
|
properties: Dict[str, Optional[str]] = {
|
|
"urlkey": None,
|
|
"timestamp": None,
|
|
"original": None,
|
|
"mimetype": None,
|
|
"statuscode": None,
|
|
"digest": None,
|
|
"length": None,
|
|
}
|
|
|
|
property_value = snapshot.split(" ")
|
|
|
|
total_property_values = len(property_value)
|
|
warranted_total_property_values = len(properties)
|
|
|
|
if total_property_values != warranted_total_property_values:
|
|
raise WaybackError(
|
|
f"Snapshot returned by CDX API has {total_property_values} prop"
|
|
f"erties instead of expected {warranted_total_property_values} "
|
|
f"properties.\nProblematic Snapshot: {snapshot}"
|
|
)
|
|
|
|
(
|
|
properties["urlkey"],
|
|
properties["timestamp"],
|
|
properties["original"],
|
|
properties["mimetype"],
|
|
properties["statuscode"],
|
|
properties["digest"],
|
|
properties["length"],
|
|
) = property_value
|
|
|
|
yield CDXSnapshot(cast(Dict[str, str], properties))
|