Use Scrapy items

docs/writing-spiders.rst

@@ -42,7 +42,7 @@ Then, for *each* URL:
 
 2. GET requests are passes it to the crawler engine to be executed. The response is passed back to the ``parse`` method in the spider.
 3. The ``parse`` method must check the return status code! If it is not 200, this must be reported by fielding a block of information.
-4. If the ``parse`` method has got a file it wants to save, it must call ``save_response_to_disk`` to do so! It should then yield a block of information.
+4. If the ``parse`` method has got a file it wants to save, it must yield something, typically using ``build_file_from_response``.
 5. The ``parse`` method can then yield further requests for processing.
 6. The blocks of information are passed to the pipeline which fires it all off to the Kingfisher Process API.
 
@@ -66,24 +66,11 @@ Here is a sample:
             # We must check the response code
             if response.status == 200:
                 # It was a success!
-                # We must call to save to the disk
-                self.save_response_to_disk(response, response.request.meta['kf_filename'])
-                # We must send some information about this success
-                yield {
-                    'success': True,
-                    'file_name': response.request.meta['kf_filename'],
-                    "data_type": "release_package",
-                    "url": response.request.url,
-                }
+                yield self.build_file_from_response(response, response.request.meta['kf_filename'], data_type='release_package')
             else:
                 # It was a failure :-(
                 # We must send some information about this failure
-                yield {
-                    'success': False,
-                    'file_name': response.request.meta['kf_filename'],
-                    "url": response.request.url,
-                    "errors": {"http_code": response.status}
-                }
+                yield self.build_file_error_from_response(response)
 
 Spider properties
 -----------------

kingfisher_scrapy/base_spider.py

@@ -9,6 +9,7 @@
 
 from kingfisher_scrapy import util
 from kingfisher_scrapy.exceptions import SpiderArgumentError
+from kingfisher_scrapy.items import File, FileError, FileItem
 
 
 class BaseSpider(scrapy.Spider):
@@ -91,38 +92,45 @@ def get_start_time(self, format):
         """
         return self.crawler.stats.get_value('start_time').strftime(format)
 
-    def save_response_to_disk(self, response, filename, data_type=None, encoding='utf-8', post_to_api=True):
+    def build_file_from_response(self, response, filename, data_type=None, encoding='utf-8', post_to_api=True):
         """
         Returns an item to yield, based on the response to a request.
         """
-        return self.save_data_to_disk(response.body, filename, response.request.url, data_type, encoding,
-                                      post_to_api)
+        return self.build_file(response.body, filename, response.request.url, data_type, encoding, post_to_api)
 
-    def save_data_to_disk(self, data, filename, url=None, data_type=None, encoding='utf-8', post_to_api=True):
+    def build_file(self, data, filename, url=None, data_type=None, encoding='utf-8', post_to_api=True):
         """
         Returns an item to yield.
         """
-        return {
-            'success': True,
+        return File({
             'file_name': filename,
             'data': data,
             'data_type': data_type,
             'url': url,
             'encoding': encoding,
             'post_to_api': post_to_api,
-        }
+        })
 
-    def _build_file_item(self, number, data, data_type, url, encoding, file_name):
-        return {
-            'success': True,
+    def build_file_item(self, number, data, data_type, url, encoding, file_name):
+        return FileItem({
             'number': number,
             'file_name': file_name,
             'data': data,
             'data_type': data_type,
             'url': url,
             'encoding': encoding,
             'post_to_api': True,
+        })
+
+    def build_file_error_from_response(self, response, **kwargs):
+        file_error = {
+            'url': response.request.url,
+            'errors': {'http_code': response.status},
         }
+        if 'kf_filename' in response.request.meta:
+            file_error['file_name'] = response.request.meta['kf_filename']
+        file_error.update(kwargs)
+        return FileError(file_error)
 
     def _get_package_metadata(self, f, skip_key):
         """
@@ -144,7 +152,7 @@ def parse_json_lines(self, f, data_type, url, encoding='utf-8', file_name='data.
                 break
             if isinstance(line, bytes):
                 line = line.decode(encoding=encoding)
-            yield self._build_file_item(number, line, data_type, url, encoding, file_name)
+            yield self.build_file_item(number, line, data_type, url, encoding, file_name)
 
     def parse_json_array(self, f_package, f_list, data_type, url, encoding='utf-8', array_field_name='releases',
                          file_name='data.json'):
@@ -158,7 +166,7 @@ def parse_json_array(self, f_package, f_list, data_type, url, encoding='utf-8',
         for number, items in enumerate(util.grouper(ijson.items(f_list, '{}.item'.format(array_field_name)), size), 1):
             package[array_field_name] = filter(None, items)
             data = json.dumps(package, default=util.default)
-            yield self._build_file_item(number, data, data_type, url, encoding, file_name)
+            yield self.build_file_item(number, data, data_type, url, encoding, file_name)
             if self.sample:
                 break
 
@@ -187,7 +195,7 @@ def parse_zipfile(self, response, data_type, file_format=None, encoding='utf-8')
         if response.status == 200:
             if file_format:
                 filename = '{}.zip'.format(hashlib.md5(response.url.encode('utf-8')).hexdigest())
-                self.save_response_to_disk(response, filename, post_to_api=False)
+                self.build_file_from_response(response, filename, post_to_api=False)
 
             zip_file = ZipFile(BytesIO(response.body))
             for finfo in zip_file.infolist():
@@ -205,15 +213,10 @@ def parse_zipfile(self, response, data_type, file_format=None, encoding='utf-8')
                     yield from self.parse_json_array(package, data, data_type, response.request.url,
                                                      encoding=encoding, file_name=filename)
                 else:
-                    yield self.save_data_to_disk(data.read(), filename, data_type=data_type, url=response.request.url,
-                                                 encoding=encoding)
+                    yield self.build_file(data.read(), filename, data_type=data_type, url=response.request.url,
+                                          encoding=encoding)
         else:
-            yield {
-                'success': False,
-                'file_name': response.request.meta['kf_filename'],
-                'url': response.request.url,
-                'errors': {'http_code': response.status}
-            }
+            yield self.build_file_error_from_response(response)
 
 
 class LinksSpider(BaseSpider):
@@ -235,14 +238,9 @@ def next_link(response):
     def parse_next_link(self, response, data_type):
         if response.status == 200:
 
-            yield self.save_response_to_disk(response, response.request.meta['kf_filename'], data_type=data_type)
+            yield self.build_file_from_response(response, response.request.meta['kf_filename'], data_type=data_type)
 
             if not self.sample:
                 yield self.next_link(response)
         else:
-            yield {
-                'success': False,
-                'file_name': response.request.meta['kf_filename'],
-                'url': response.request.url,
-                'errors': {'http_code': response.status}
-            }
+            yield self.build_file_error_from_response(response)

kingfisher_scrapy/extensions.py

@@ -1,13 +1,15 @@
+# https://docs.scrapy.org/en/latest/topics/extensions.html#writing-your-own-extension
+
 import json
 import os
 
 from scrapy import signals
 from scrapy.exceptions import NotConfigured
 
+from kingfisher_scrapy.items import File, FileError, FileItem
 from kingfisher_scrapy.kingfisher_process import Client
 
 
-# https://docs.scrapy.org/en/latest/topics/extensions.html#writing-your-own-extension
 class KingfisherFilesStore:
     def __init__(self, directory):
         self.directory = directory
@@ -21,12 +23,11 @@ def from_crawler(cls, crawler):
 
     def item_scraped(self, item, spider):
         """
-        Writes the item's data to the filename in the crawl's directory.
+        If the item is a file, writes its data to the filename in the crawl's directory.
 
         Writes a ``<filename>.fileinfo`` metadata file in the crawl's directory, and returns a dict with the metadata.
         """
-        # Skip failures and parts of files.
-        if not item['success'] or 'number' in item:
+        if not isinstance(item, File):
             return
 
         # The crawl's relative directory, in the format `<spider_name>[_sample]/<YYMMDD_HHMMSS>`.
@@ -119,14 +120,18 @@ def item_scraped(self, item, spider):
             'url': item['url'],
         }
 
-        if item['success']:
+        if isinstance(item, FileError):
+            data['errors'] = json.dumps(item['errors'])
+
+            self._request(item, spider, 'create_file_error', data, name='File Errors API')
+        else:
             data['data_type'] = item['data_type']
             data['encoding'] = item.get('encoding', 'utf-8')
             if spider.note:
                 data['collection_note'] = spider.note
 
             # File Item
-            if 'number' in item:
+            if isinstance(item, FileItem):
                 data['number'] = item['number']
                 data['data'] = item['data']
 
@@ -145,12 +150,6 @@ def item_scraped(self, item, spider):
 
                 self._request(item, spider, 'create_file', data, files)
 
-        # File Error
-        else:
-            data['errors'] = json.dumps(item['errors'])
-
-            self._request(item, spider, 'create_file_error', data, name='File Errors API')
-
     def _request(self, item, spider, method, *args, name='API'):
         response = getattr(self.client, method)(*args)
         if not response.ok:

kingfisher_scrapy/items.py

@@ -1,14 +1,30 @@
-# -*- coding: utf-8 -*-
+import scrapy
 
-# Define here the models for your scraped items
-#
-# See documentation in:
-# https://docs.scrapy.org/en/latest/topics/items.html
 
-import scrapy
+class File(scrapy.Item):
+    file_name = scrapy.Field()
+    url = scrapy.Field()
+    data = scrapy.Field()
+    data_type = scrapy.Field()
+    encoding = scrapy.Field()
+    post_to_api = scrapy.Field()
+
+    # Added by extensions.
+    path = scrapy.Field()
+    files_store = scrapy.Field()
+
+
+class FileItem(scrapy.Item):
+    number = scrapy.Field()
+    file_name = scrapy.Field()
+    url = scrapy.Field()
+    data = scrapy.Field()
+    data_type = scrapy.Field()
+    encoding = scrapy.Field()
+    post_to_api = scrapy.Field()
 
 
-class KingfisherScrapyItem(scrapy.Item):
-    # define the fields for your item here like:
-    # name = scrapy.Field()
-    pass
+class FileError(scrapy.Item):
+    file_name = scrapy.Field()
+    url = scrapy.Field()
+    errors = scrapy.Field()

kingfisher_scrapy/spiders/afghanistan_records.py

@@ -32,17 +32,12 @@ def parse_list(self, response):
                     callback=self.parse_record
                 )
         else:
-            yield {
-                'success': False,
-                'file_name': 'list.json',
-                "url": response.request.url,
-                "errors": {"http_code": response.status}
-            }
+            yield self.build_file_error_from_response(response, file_name='list.json')
 
     def parse_record(self, response):
         if response.status == 200:
 
-            yield self.save_response_to_disk(response, response.request.meta['kf_filename'], data_type="record")
+            yield self.build_file_from_response(response, response.request.meta['kf_filename'], data_type="record")
 
         elif response.status == 429:
             self.crawler.engine.pause()
@@ -51,15 +46,10 @@ def parse_record(self, response):
             url = response.request.url
             # This is dangerous as we might get stuck in a loop here if we always get a 429 response. Try this for now.
             yield scrapy.Request(
-                    url=url,
-                    meta={'kf_filename': url.split('/')[-1]+'.json'},
-                    callback=self.parse_record,
-                    dont_filter=True,
-                )
+                url=url,
+                meta={'kf_filename': url.split('/')[-1]+'.json'},
+                callback=self.parse_record,
+                dont_filter=True,
+            )
         else:
-            yield {
-                'success': False,
-                'file_name': response.request.meta['kf_filename'],
-                "url": response.request.url,
-                "errors": {"http_code": response.status}
-            }
+            yield self.build_file_error_from_response(response)

kingfisher_scrapy/spiders/afghanistan_releases.py

@@ -31,12 +31,7 @@ def parse_list(self, response):
                     callback=self.parse_release_list
                 )
         else:
-            yield {
-                'success': False,
-                'file_name': 'list.json',
-                "url": response.request.url,
-                "errors": {"http_code": response.status}
-            }
+            yield self.build_file_error_from_response(response, filename='list.json')
 
     def parse_release_list(self, response):
         if response.status == 200:
@@ -64,17 +59,12 @@ def parse_release_list(self, response):
                 dont_filter=True,
             )
         else:
-            yield {
-                'success': False,
-                'file_name': response.request.meta['kf_filename'],
-                "url": response.request.url,
-                "errors": {"http_code": response.status}
-            }
+            yield self.build_file_error_from_response(response)
 
     def parse_release(self, response):
         if response.status == 200:
 
-            yield self.save_response_to_disk(response, response.request.meta['kf_filename'], data_type="release")
+            yield self.build_file_from_response(response, response.request.meta['kf_filename'], data_type="release")
 
         elif response.status == 429:
             self.crawler.engine.pause()
@@ -89,9 +79,4 @@ def parse_release(self, response):
                 dont_filter=True,
             )
         else:
-            yield {
-                'success': False,
-                'file_name': response.request.meta['kf_filename'],
-                "url": response.request.url,
-                "errors": {"http_code": response.status}
-            }
+            yield self.build_file_error_from_response(response)

kingfisher_scrapy/spiders/argentina_buenos_aires.py

@@ -33,12 +33,7 @@ def parse_list(self, response):
                 if resource['format'].upper() == 'JSON':
                     yield scrapy.Request(url=resource['url'])
         else:
-            yield {
-                'success': False,
-                'file_name': 'list.json',
-                'url': response.request.url,
-                'errors': {'http_code': response.status}
-            }
+            yield self.build_file_error_from_response(response, filename='list.json')
 
     def parse(self, response):
         yield from self.parse_zipfile(response, 'release_package', file_format='release_package')

kingfisher_scrapy/spiders/argentina_vialidad.py

@@ -13,11 +13,6 @@ def start_requests(self):
 
     def parse(self, response):
         if response.status == 200:
-            yield self.save_response_to_disk(response, 'all.json', data_type='release_package_list')
+            yield self.build_file_from_response(response, 'all.json', data_type='release_package_list')
         else:
-            yield {
-                'success': False,
-                'file_name': 'all.json',
-                "url": response.request.url,
-                "errors": {"http_code": response.status}
-            }
+            yield self.build_file_error_from_response(response, filename='all.json')