def main():
"""Main worker entrypoint."""
LOG.warning("worker online")
# until we can resume jobs
for state in (Keys.processing.value, Keys.pending.value):
CACHE.delete_many(*utils.list_keys(state))
while True:
prune = []
for glet in WORKERS:
if glet.successful():
prune.append(glet)
elif glet.dead:
LOG.warning(
"worker crashed: %s",
"".join(format_exception(*glet.exc_info)).strip(),
)
prune.append(glet)
for glet in prune:
WORKERS.remove(glet)
process_new()
gc.collect()
gevent.sleep(10)
def write_data(uuid, data):
"""Try to store the data, log errors."""
try:
_write_data(uuid, data)
except Exception as error:
LOG.warning("Failed to save data: %r", error)
def knife(uuid, token, verify, roles): # pylint: disable=R0914
"""Pull all ESI data for a character_id.
Args:
uuid: string uuid token
token: SSO access token
verify: dictionary return from /verify/
roles: list of corporation roles
"""
character_id = verify["CharacterID"]
LOG.warning("knife run started for character: %s", character_id)
scopes = verify["Scopes"]
_, _, public = utils.request_or_wait("{}/latest/characters/{}/".format(
ESI, character_id))
if isinstance(public, str):
CACHE.delete("{}{}".format(Keys.processing.value, uuid))
utils.write_data(uuid, {"public info failure": public})
return
headers = {"Authorization": "Bearer {}".format(token)}
results = get_results(public, character_id, scopes, roles, headers)
utils.write_data(uuid, results)
CACHE.delete("{}{}".format(Keys.processing.value, uuid))
CACHE.cache.inc(Keys.alltime.value, 1)
LOG.warning("completed character: %r", character_id)
def request_or_wait(url,
*args,
_as_res=False,
page=None,
method="get",
**kwargs):
"""Request the URL, or wait if we're error limited."""
check_x_pages = True
if page:
kwargs["params"] = kwargs.get("params", {})
kwargs["params"]["page"] = page
check_x_pages = False
LOG.warning("requesting: %s (page %d)", url, page)
else:
LOG.warning("requesting: %s", url)
try:
res = getattr(SESSION, method)(url, *args, **kwargs)
res.raise_for_status()
except Exception as err:
try:
if res.status_code == 420:
wait = int(res.headers.get("X-Esi-Error-Limit-Reset", 1)) + 1
APP.error_limited = True
LOG.warning("hit the error limit, waiting %d seconds", wait)
# error limited. wait out the window then carry on
gevent.sleep(wait)
return request_or_wait(url,
*args,
_as_res=_as_res,
page=page,
method=method,
**kwargs)
except Exception as error:
LOG.warning("error handling error: %r: %r", err, error)
try:
content = res.json()
except Exception:
content = res.text
# /shrug some other error, can't win em all
return None, url, \
res if _as_res else "Error fetching data: {} {}".format(
res.status_code,
content,
)
else:
if check_x_pages:
try:
pages = list(range(2, int(res.headers.get("X-Pages", 0))))
except Exception as error:
LOG.warning("error checking x-pages for %s: %r", url, error)
pages = None
else:
pages = page
return pages, url, res if _as_res else res.json()
def process_new():
"""Process all new tokens, verify or we're done early."""
for new_key in utils.list_keys(Keys.new.value):
uuid = new_key.split(".")[-1]
LOG.warning("processing new uuid: %r", uuid)
token = CACHE.get(new_key)
CACHE.delete(new_key)
if not token:
LOG.warning("no token stored for uuid: %r", uuid)
continue
pending_key = "{}{}".format(Keys.pending.value, uuid)
CACHE.set(
pending_key,
"1",
timeout=70,
)
headers = {"Authorization": "Bearer {}".format(token)}
_, res = utils.request_or_wait(
"{}/verify/".format(ESI),
headers=headers,
)
failed = False
if isinstance(res, str) or "CharacterID" not in res:
utils.write_data(uuid, {"auth failure": res})
failed = True
else:
_, roles = utils.request_or_wait(
"{}/latest/characters/{}/roles/".format(
ESI,
res["CharacterID"],
),
headers=headers,
)
if isinstance(roles, str):
utils.write_data(uuid, {"roles failure": roles})
failed = True
CACHE.delete(pending_key)
if not failed:
CACHE.set(
"{}{}".format(Keys.processing.value, uuid),
res["CharacterID"],
timeout=7200,
)
WORKERS.append(
gevent.spawn(knife, uuid, token, res, roles)
)
def write_data(uuid, data):
"""Try to store the data, log errors."""
try:
CACHE.set(
"{}{}".format(Keys.complete.value, uuid),
codecs.decode(
base64.b64encode(
gzip.compress(codecs.encode(
ujson.dumps(data),
"utf-8",
))),
"utf-8",
),
timeout=EXPIRY,
)
except Exception as error:
LOG.warning("Failed to save data: %r", error)
def get_data(filename, decompress=True):
"""Open and return the content from file."""
filepath = os.path.join(DATA, filename)
try:
with open(filepath, "r") as openfile:
content = openfile.read()
except Exception as error:
LOG.warning("failed to open %s: %r", filename, error)
else:
try:
if decompress:
return ujson.loads(gzip.decompress(base64.b64decode(content)))
return ujson.loads(content)
except Exception as error:
LOG.warning("failed to decode %s: %r", filename, error)
return None
def _get_names(ids):
"""Resolve ids to names."""
resolved = {}
failed = []
for i in range(0, len(ids), 1000):
batch = ids[i:i+1000]
_, _, res = utils.request_or_wait(
"{}/latest/universe/names/".format(ESI),
method="post",
json=batch,
)
if isinstance(res, list):
for _res in res:
resolved[_res["id"]] = _res["name"]
else:
failed.extend(batch)
while failed:
still_failed = []
random.shuffle(failed)
batch_size = max(min(int(len(failed) / 2), 500), 1)
for i in range(0, len(failed), batch_size):
batch = failed[i:i+batch_size]
_, _, res = utils.request_or_wait(
"{}/latest/universe/names/".format(ESI),
method="post",
json=batch,
)
if isinstance(res, list):
for _res in res:
resolved[_res["id"]] = _res["name"]
else:
still_failed.extend(batch)
failed = still_failed
if batch_size == 1 and still_failed:
LOG.warning("failed to resolve: %r", still_failed)
break
return resolved
def refresh_spec():
"""Refresh the ESI spec.
Returns:
dictionary: JSON loaded swagger spec
"""
try:
spec_details = CACHE.get(Keys.spec.value)
except redis.exceptions.ConnectionError:
spec_details = None
save_results = False
else:
save_results = True
if spec_details is None:
spec_details = {"timestamp": 0}
if time.time() - spec_details["timestamp"] > 300:
headers = {}
if spec_details.get("etag"):
headers["If-None-Match"] = spec_details["etag"]
_, _, res = request_or_wait(
"{}/latest/swagger.json".format(ESI),
_as_res=True,
headers=headers,
)
if isinstance(res, str):
LOG.warning("failed to refresh spec: %s", res)
return spec_details.get("spec", {})
spec_details["timestamp"] = time.time()
if res.status_code != 304:
spec_details["etag"] = res.headers.get("ETag")
spec_details["spec"] = JsonDeref().deref(res.json())
if save_results:
CACHE.set(Keys.spec.value, spec_details, timeout=3600)
return spec_details["spec"]
def request_or_wait(url, *args, _as_res=False, **kwargs):
"""Request the URL, or wait if we're error limited."""
try:
LOG.warning("requesting: %s", url)
res = SESSION.get(url, *args, **kwargs)
res.raise_for_status()
return url, res if _as_res else res.json()
except Exception as err:
try:
if res.status_code == 420:
wait = int(res.headers.get("X-Esi-Error-Limit-Reset", 1)) + 1
LOG.warning("hit the error limit, waiting %d seconds", wait)
# error limited. wait out the window then carry on
gevent.sleep(wait)
return request_or_wait(url, *args, _as_res=_as_res, **kwargs)
except Exception as error:
LOG.warning("error handling error: %r: %r", err, error)
try:
content = res.json()
except Exception:
content = res.text
# /shrug some other error, can't win em all
return url, res if _as_res else "Error fetching data: {} {}".format(
res.status_code,
content,
)
def refresh_spec():
"""Refresh the ESI spec.
Returns:
dictionary: JSON loaded swagger spec
"""
spec_file = os.path.join(DATA, ".esi.json")
if os.path.isfile(spec_file):
with open(spec_file, "r") as open_spec:
spec_details = ujson.loads(open_spec.read())
else:
spec_details = {"timestamp": 0}
if time.time() - spec_details["timestamp"] > 300:
headers = {}
if spec_details.get("etag"):
headers["If-None-Match"] = spec_details["etag"]
_, res = request_or_wait(
"{}/latest/swagger.json".format(ESI),
_as_res=True,
headers=headers,
)
if isinstance(res, str):
LOG.warning("failed to refresh spec: %s", res)
return spec_details.get("spec", {})
spec_details["timestamp"] = time.time()
if res.status_code != 304:
spec_details["etag"] = res.headers.get("ETag")
spec_details["spec"] = JsonDeref().deref(res.json())
with open(spec_file, "w") as new_spec:
new_spec.write(ujson.dumps(spec_details))
return spec_details["spec"]
def get_data(uuid):
"""Open and return the character's data."""
cache_key = "{}{}".format(Keys.complete.value, uuid)
try:
content = CACHE.get(cache_key)
except Exception as error:
LOG.warning("failed to get %s: %r", cache_key, error)
else:
if content is None:
return None
try:
return ujson.loads(gzip.decompress(base64.b64decode(content)))
except Exception as error:
LOG.warning("failed to decode %s: %r", content, error)
else:
CACHE.cache._client.expire( # pylint: disable=protected-access
cache_key,
EXPIRY,
)
return None
def knife(uuid, token, verify, roles): # pylint: disable=R0914
"""Pull all ESI data for a character_id.
Args:
uuid: string uuid token
token: SSO access token
verify: dictionary return from /verify/
roles: list of corporation roles
"""
character_id = verify["CharacterID"]
LOG.warning("knife run started for character: %s", character_id)
scopes = verify["Scopes"]
_, public = utils.request_or_wait(
"{}/latest/characters/{}/".format(ESI, character_id)
)
if isinstance(public, str):
CACHE.delete("{}{}".format(Keys.processing.value, uuid))
utils.write_data(uuid, {"public info failure": public})
return
all_params = copy.deepcopy(ADDITIONAL_PARAMS)
known_params = {"character_id": character_id}
if public["corporation_id"] > 2000000:
known_params["corporation_id"] = public["corporation_id"]
else:
all_params.pop("corporation_id")
if "alliance_id" in public:
known_params["alliance_id"] = public["alliance_id"]
spec = utils.refresh_spec()
headers = {"Authorization": "Bearer {}".format(token)}
results = expand_params(
scopes,
roles,
spec,
known_params,
all_params,
headers,
)
urls = build_urls(scopes, roles, spec, known_params, all_params)
with ThreadPoolExecutor(max_workers=20) as pool:
futures = []
for url in urls:
futures.append(pool.submit(
utils.request_or_wait,
url,
headers=headers,
))
for future in as_completed(futures):
url, result = future.result()
results[url] = result
utils.write_data(uuid, results)
CACHE.delete("{}{}".format(Keys.processing.value, uuid))
LOG.warning("completed character: %r", character_id)
def expand_params(scopes, roles, spec, # pylint: disable=R0914,R0913
known_params, all_params, headers):
"""Gather IDs from all_params into known_params."""
errors = []
purge = {x: [] for x in all_params}
transform = {
"/characters/{character_id}/mail/labels/": \
lambda x: [i["label_id"] for i in x["labels"]],
"/characters/{character_id}/planets/": \
lambda x: [i["planet_id"] for i in x],
"/characters/{character_id}/calendar/": \
lambda x: [i["event_id"] for i in x],
"/characters/{character_id}/contracts/": \
lambda x: [i["contract_id"] for i in x],
"/corporations/{corporation_id}/calendar/": \
lambda x: [i["event_id"] for i in x],
"/corporations/{corporation_id}/contracts/": \
lambda x: [i["contract_id"] for i in x],
}
expansion_results = {}
with ThreadPoolExecutor(max_workers=20) as pool:
futures = {}
for parent, id_types in all_params.items():
for id_type, url in id_types.items():
oper = spec["paths"][url]["get"]
required_roles = oper.get("x-required-roles", [])
if any(x not in roles for x in required_roles):
# we don't have the corporate roles for this route
purge[parent].append(id_type)
continue
required_sso = oper.get("security", [{}])[0].get("evesso", [])
if any(x not in scopes for x in required_sso):
# our access token doesn't have this scope
purge[parent].append(id_type)
continue
path = "https://esi.evetech.net/latest{}".format(
url.format(**known_params)
)
futures[pool.submit(
utils.request_or_wait,
path,
headers=headers,
)] = (url, parent, id_type)
for future in as_completed(futures):
templated_url, parent, id_type = futures[future]
url, data = future.result()
if templated_url in transform:
expansion_results[url] = data
all_params[parent][id_type] = transform[templated_url](data)
elif isinstance(data, list):
all_params[parent][id_type] = data
else:
LOG.warning("worker expansion error: %r", data)
for parent, purged_ids in purge.items():
for purged_id in purged_ids:
all_params[parent].pop(purged_id)
if errors:
LOG.warning("worker errors: %s", " ".join(errors))
return expansion_results
def expand_params(scopes, roles, spec, # pylint: disable=R0914,R0913
known_params, all_params, headers):
"""Gather IDs from all_params into known_params."""
errors = []
purge = {x: [] for x in all_params}
transform = {
"/characters/{character_id}/mail/labels/": \
lambda x: [i["label_id"] for i in x["labels"]],
"/characters/{character_id}/planets/": \
lambda x: [i["planet_id"] for i in x],
"/characters/{character_id}/calendar/": \
lambda x: [i["event_id"] for i in x],
"/characters/{character_id}/contracts/": \
lambda x: [i["contract_id"] for i in x],
"/characters/{character_id}/mail/": \
lambda x: [i["mail_id"] for i in x],
"/corporations/{corporation_id}/calendar/": \
lambda x: [i["event_id"] for i in x],
"/corporations/{corporation_id}/contracts/": \
lambda x: [i["contract_id"] for i in x],
}
expansion_results = {}
with ThreadPoolExecutor(max_workers=20) as pool:
futures = {}
for parent, id_types in all_params.items():
for id_type, url in id_types.items():
oper = spec["paths"][url]["get"]
required_roles = oper.get("x-required-roles", [])
if any(x not in roles for x in required_roles):
# we don't have the corporate roles for this route
purge[parent].append(id_type)
continue
required_sso = oper.get("security", [{}])[0].get("evesso", [])
if any(x not in scopes for x in required_sso):
# our access token doesn't have this scope
purge[parent].append(id_type)
continue
path = "https://esi.evetech.net/latest{}".format(
url.format(**known_params)
)
futures[pool.submit(
utils.request_or_wait,
path,
headers=headers,
)] = (url, parent, id_type)
pages = {}
while True:
completed = []
expansion = {}
for future in as_completed(futures):
completed.append(future)
templated_url, parent, id_type = futures[future]
page, url, data = future.result()
page_key = (templated_url, parent, id_type, url)
if page and isinstance(page, list):
pages[page_key] = {1: data}
for _page in page:
expansion[pool.submit(
utils.request_or_wait,
url,
page=_page,
headers=headers,
)] = (templated_url, parent, id_type)
elif isinstance(page, int):
if isinstance(data, list):
pages[page_key][page] = data
else:
LOG.warning("worker page expansion error: %r", data)
else:
if templated_url in transform:
expansion_results[url] = data
all_params[parent][id_type] = transform[templated_url](
data
)
elif isinstance(data, list):
all_params[parent][id_type] = data
else:
LOG.warning("worker expansion error: %r", data)
for complete in completed:
futures.pop(complete)
futures.update(expansion)
if not futures:
break
for details, page_data in pages.items():
templated_url, parent, id_type, url = details
data = []
for page in sorted(page_data):
data.extend(page_data[page])
if not data:
continue
if templated_url in transform:
expansion_results[url] = data
try:
all_params[parent][id_type] = transform[templated_url](
data
)
except Exception as error:
LOG.warning(
"failed to transform %s. error: %r data: %r",
url,
error,
data,
)
else:
all_params[parent][id_type] = data
for parent, purged_ids in purge.items():
for purged_id in purged_ids:
all_params[parent].pop(purged_id)
if errors:
LOG.warning("worker errors: %s", " ".join(errors))
return expansion_results
