Merge pull request #35 from Sefaria/dev

Dev

api/api_warnings.py

@@ -0,0 +1,62 @@
+import django
+django.setup()
+from sefaria.model import *
+from typing import List
+from enum import Enum
+
+class APIWarningCode(Enum):
+    APINoVersion = 101
+    APINoLanguageVersion = 102
+    APINoSourceText = 103
+    APINoTranslationText = 104
+
+"""
+classes for data warnings in API calls.
+used when part of the data that was requested exists and returned, and part is missing.  
+"""
+
+class APIDatawarning():
+    """
+    general class
+    """
+
+    def __init__(self):
+        pass
+
+
+class TextsAPIResponseMessage(APIDatawarning):
+    """
+    class for returning a message and an warning code
+    """
+
+    def get_message(self) -> dict:
+        return {'warning_code': self.warning_code.value,
+                'message': self.message}
+
+
+class APINoVersion(TextsAPIResponseMessage):
+
+    def __init__(self, oref: Ref, vtitle: str, lang: str):
+        self.warning_code = APIWarningCode.APINoVersion
+        self.message = f'We do not have version named {vtitle} with language {lang} for {oref}'
+
+
+class APINoLanguageVersion(TextsAPIResponseMessage):
+
+    def __init__(self, oref: Ref, langs: List[str]):
+        self.warning_code = APIWarningCode.APINoLanguageVersion
+        self.message = f'We do not have the language you asked for {oref}. Available languages are {langs}'
+
+
+class APINoSourceText(TextsAPIResponseMessage):
+
+    def __init__(self, oref: Ref):
+        self.warning_code = APIWarningCode.APINoSourceText
+        self.message = f'We do not have the source text for {oref}'
+
+
+class APINoTranslationText(TextsAPIResponseMessage):
+
+    def __init__(self, oref: Ref):
+        self.warning_code = APIWarningCode.APINoTranslationText
+        self.message = f'We do not have a translation for {oref}'

api/tests.py

@@ -0,0 +1,218 @@
+from django.test.client import Client
+import django
+django.setup()
+from reader.tests import SefariaTestCase
+import json
+from api.api_warnings import APIWarningCode
+
+
+c = Client()
+
+
+class APITextsTests(SefariaTestCase):
+
+    def test_api_get_text_default(self):
+        response = c.get('/api/v3/texts/Genesis.1')
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertTrue(len(data["versions"]) == 1)
+        self.assertTrue(data["versions"][0]['actualLanguage'] == 'he')
+        self.assertEqual(data["book"], "Genesis")
+        self.assertEqual(data["categories"], ["Tanakh", "Torah"])
+        self.assertEqual(data["sections"], ['1'])
+        self.assertEqual(data["toSections"], ['1'])
+
+    def test_api_get_text_source_all(self):
+        response = c.get('/api/v3/texts/Shabbat.22a?version=source|all')
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertTrue(len(data["versions"]) > 1)
+        self.assertTrue(all(v['actualLanguage'] == 'he' for v in data["versions"]))
+        self.assertEqual(data["book"], "Shabbat")
+        self.assertEqual(data["categories"], ["Talmud", "Bavli", "Seder Moed"])
+        self.assertEqual(data["sections"], ["22a"])
+        self.assertEqual(data["toSections"], ["22a"])
+
+    def test_api_get_text_source(self):
+        response = c.get('/api/v3/texts/Shabbat.22a?version=source')
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(len(data["versions"]), 1)
+        self.assertEqual(data["versions"][0]['versionTitle'], "William Davidson Edition - Vocalized Aramaic")
+
+    def test_api_get_text_translation_all(self):
+        response = c.get('/api/v3/texts/Shabbat.22a?version=translation|all')
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertTrue(len(data["versions"]) > 1)
+        self.assertTrue(any(v['actualLanguage'] == 'en' for v in data["versions"]))
+        self.assertEqual(data["book"], "Shabbat")
+        self.assertEqual(data["categories"], ["Talmud", "Bavli", "Seder Moed"])
+        self.assertEqual(data["sections"], ["22a"])
+        self.assertEqual(data["toSections"], ["22a"])
+
+    def test_api_get_text_translation(self):
+        response = c.get('/api/v3/texts/Shabbat.22a?version=translation')
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(len(data["versions"]), 1)
+        self.assertEqual(data["versions"][0]['versionTitle'], "William Davidson Edition - English")
+
+    def test_api_get_text_lang_all(self):
+        response = c.get('/api/v3/texts/Rashi_on_Genesis.2.3?version=english|all')
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertTrue(len(data["versions"]) > 1)
+        self.assertTrue(all(v['actualLanguage'] == 'en' for v in data["versions"]))
+        self.assertEqual(data["book"], "Rashi on Genesis")
+        self.assertEqual(data["collectiveTitle"], "Rashi")
+        self.assertEqual(data["categories"], ["Tanakh", "Rishonim on Tanakh", "Rashi", "Torah"])
+        self.assertEqual(data["sections"], ['2', '3'])
+        self.assertEqual(data["toSections"], ['2', '3'])
+
+    def test_api_get_text_specific(self):
+        response = c.get('/api/v3/texts/Tosafot_on_Sukkah.2a.4.1?version=hebrew|Vilna_Edition')
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(len(data["versions"]), 1)
+        self.assertEqual(data["versions"][0]['actualLanguage'], 'he')
+        self.assertEqual(data["versions"][0]['versionTitle'], 'Vilna Edition')
+        self.assertEqual(data["book"], "Tosafot on Sukkah")
+        self.assertEqual(data["collectiveTitle"], "Tosafot")
+        self.assertEqual(data["categories"], ["Talmud", "Bavli", "Rishonim on Talmud", "Tosafot", "Seder Moed"])
+        self.assertEqual(data["sections"], ["2a", '4', '1'])
+        self.assertEqual(data["toSections"], ["2a", '4', '1'])
+
+    def test_api_get_text_primary_all(self):
+        response = c.get('/api/v3/texts/Genesis.1?version=primary|all')
+        data = json.loads(response.content)
+        self.assertTrue(len(data["versions"]) > 3)
+        self.assertTrue(all(v['actualLanguage'] == 'he' for v in data["versions"]))
+
+    def test_api_get_text_primary(self):
+        response = c.get('/api/v3/texts/Shabbat.22a?version=primary')
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(len(data["versions"]), 1)
+        self.assertEqual(data["versions"][0]['versionTitle'], "William Davidson Edition - Vocalized Aramaic")
+
+    def test_api_get_text_two_params(self):
+        response = c.get('/api/v3/texts/Genesis.1?version=hebrew|Tanach with Nikkud&version=english|all')
+        data = json.loads(response.content)
+        self.assertTrue(len(data["versions"]) > 7)
+        self.assertEqual(data["versions"][0]['actualLanguage'], 'he')
+        self.assertTrue(all(v['actualLanguage'] == 'en' for v in data["versions"][1:]))
+
+    def test_api_get_text_range(self):
+        response = c.get('/api/v3/texts/Job.5.2-4')
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(data["sections"], ['5', '2'])
+        self.assertEqual(data["toSections"], ['5', '4'])
+
+    def text_api_virtual_node(self):
+        response = c.get('/api/v3/texts/BDB, א')
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(len(data['versions']), 1)
+        self.assertEqual(data['versions'][0]['text'], ['<big><span dir="rtl">א</span></big>  <em>Āleph</em>, first letter; in post Biblical Hebrew = numeral 1 (and so in margin of printed MT); א̈= 1000; no evidence of this usage in OT times.'])
+
+    def test_api_get_text_bad_text(self):
+        response = c.get('/api/v3/texts/Life_of_Pi.13.13')
+        self.assertEqual(404, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(data["error"], "Could not find title in reference: Life of Pi.13.13")
+
+    def test_api_get_text_out_of_bound(self):
+        response = c.get('/api/v3/texts/Genesis.999')
+        data = json.loads(response.content)
+        self.assertEqual(data["error"], "Genesis ends at Chapter 50.")
+
+    def test_api_get_text_too_many_hyphens(self):
+        response = c.get('/api/v3/texts/Genesis.9-4-5')
+        data = json.loads(response.content)
+        self.assertEqual(data["error"], "Couldn't understand ref 'Genesis.9-4-5' (too many -'s).")
+
+    def test_api_get_text_bad_sections(self):
+        response = c.get('/api/v3/texts/Job.6-X')
+        self.assertEqual(404, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(data["error"], "Couldn't understand text sections: 'Job.6-X'.")
+
+    def test_api_get_text_empty_ref(self):
+        response = c.get("/api/v3/texts/Berakhot.1a")
+        self.assertEqual(404, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(data["error"], "We have no text for Berakhot 1a.")
+
+    def test_api_get_text_no_source(self):
+        response = c.get("/api/v3/texts/The_Book_of_Maccabees_I.1?version=english|Brenton's_Septuagint&version=source")
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(len(data["versions"]), 1)
+        self.assertEqual(data['warnings'][0]['source']['warning_code'], APIWarningCode.APINoSourceText.value)
+        self.assertEqual(data['warnings'][0]['source']['message'], 'We do not have the source text for The Book of Maccabees I 1')
+
+    def test_api_get_text_no_translation(self):
+        response = c.get("/api/v3/texts/Shuvi_Shuvi_HaShulamit?version=translation")
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(len(data["versions"]), 0)
+        self.assertEqual(data['warnings'][0]['translation']['warning_code'], APIWarningCode.APINoTranslationText.value)
+        self.assertEqual(data['warnings'][0]['translation']['message'], 'We do not have a translation for Shuvi Shuvi HaShulamit')
+
+    def test_api_get_text_no_language(self):
+        response = c.get("/api/v3/texts/The_Book_of_Maccabees_I.1?version=english|Brenton's_Septuagint&version=sgrg|all")
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(len(data["versions"]), 1)
+        self.assertEqual(data['warnings'][0]['sgrg|all']['warning_code'], APIWarningCode.APINoLanguageVersion.value)
+        self.assertEqual(data['warnings'][0]['sgrg|all']['message'],
+                         "We do not have the language you asked for The Book of Maccabees I 1. Available languages are ['english', 'hebrew']")
+
+    def test_api_get_text_no_version(self):
+        response = c.get("/api/v3/texts/The_Book_of_Maccabees_I.1?version=english|Brenton's_Septuagint&version=hebrew|Kishkoosh")
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(len(data["versions"]), 1)
+        self.assertEqual(data['warnings'][0]['hebrew|Kishkoosh']['warning_code'], APIWarningCode.APINoVersion.value)
+        self.assertEqual(data['warnings'][0]['hebrew|Kishkoosh']['message'],
+                         'We do not have version named Kishkoosh with language hebrew for The Book of Maccabees I 1')
+
+    def test_fill_in_missing_segments(self):
+        vtitle = "Maimonides' Mishneh Torah, edited by Philip Birnbaum, New York, 1967"
+        response = c.get(f"/api/v3/texts/Mishneh_Torah,_Sabbath_1?version=english|{vtitle}&fill_in_missing_segments=true")
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertTrue(len(data['versions'][0]['text']) > 2)
+        self.assertTrue(data['versions'][0].get('sources'))
+        self.assertEqual(data['versions'][0]['sources'][0], vtitle)
+        self.assertNotEqual(data['versions'][0]['sources'][2], vtitle)
+
+    def test_without_fill_in_missing_segments(self):
+        vtitle = "Maimonides' Mishneh Torah, edited by Philip Birnbaum, New York, 1967"
+        response = c.get(f"/api/v3/texts/Mishneh_Torah,_Sabbath_1?version=english|{vtitle}")
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(len(data['versions'][0]['text']), 2)
+        self.assertFalse(data['versions'][0].get('sources'))
+
+    def test_wrap_all_entities(self):
+        vtitle = "The Contemporary Torah, Jewish Publication Society, 2006"
+        response = c.get(f"/api/v3/texts/Genesis%2010?version=english|{vtitle}&return_format=wrap_all_entities")
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertTrue('<a class ="refLink"' in data['versions'][0]['text'][3])
+        self.assertTrue('<a href="/topics' in data['versions'][0]['text'][8])
+
+    def test_text_only(self):
+        response = c.get(f"/api/v3/texts/Shulchan_Arukh%2C_Orach_Chayim.1:1?return_format=text_only")
+        self.assertEqual(200, response.status_code)
+        data = json.loads(response.content)
+        self.assertFalse('<' in data['versions'][0]['text'])
+
+    def test_error_return_format(self):
+        response = c.get(f"/api/v3/texts/Shulchan_Arukh%2C_Orach_Chayim.1:1?return_format=not_valid")
+        self.assertEqual(400, response.status_code)
+        data = json.loads(response.content)
+        self.assertEqual(data['error'], "return_format should be one of those formats: ['default', 'wrap_all_entities', 'text_only'].")

api/views.py

@@ -0,0 +1,58 @@
+from sefaria.model import *
+from sefaria.model.text_reuqest_adapter import TextRequestAdapter
+from sefaria.client.util import jsonResponse
+from django.views import View
+from .api_warnings import *
+
+
+class Text(View):
+
+    RETURN_FORMATS = ['default', 'wrap_all_entities', 'text_only']
+
+    def dispatch(self, request, *args, **kwargs):
+        try:
+            self.oref = Ref.instantiate_ref_with_legacy_parse_fallback(kwargs['tref'])
+        except Exception as e:
+            return jsonResponse({'error': getattr(e, 'message', str(e))}, status=404)
+        return super().dispatch(request, *args, **kwargs)
+
+    @staticmethod
+    def split_piped_params(params_string) -> List[str]:
+        params = params_string.split('|')
+        if len(params) < 2:
+            params.append('')
+        params[1] = params[1].replace('_', ' ')
+        return params
+
+    def _handle_warnings(self, data):
+        data['warnings'] = []
+        for lang, vtitle in data['missings']:
+            if lang == 'source':
+                warning = APINoSourceText(self.oref)
+            elif lang == 'translation':
+                warning = APINoTranslationText(self.oref)
+            elif vtitle and vtitle != 'all':
+                warning = APINoVersion(self.oref, vtitle, lang)
+            else:
+                warning = APINoLanguageVersion(self.oref, data['available_langs'])
+            representing_string = f'{lang}|{vtitle}' if vtitle else lang
+            data['warnings'].append({representing_string: warning.get_message()})
+        data.pop('missings')
+        data.pop('available_langs')
+        return data
+
+    def get(self, request, *args, **kwargs):
+        if self.oref.is_empty() and not self.oref.index_node.is_virtual:
+            return jsonResponse({'error': f'We have no text for {self.oref}.'}, status=404)
+        versions_params = request.GET.getlist('version', [])
+        if not versions_params:
+            versions_params = ['primary']
+        versions_params = [self.split_piped_params(param_str) for param_str in versions_params]
+        fill_in_missing_segments = request.GET.get('fill_in_missing_segments', False)
+        return_format = request.GET.get('return_format', 'default')
+        if return_format not in self.RETURN_FORMATS:
+            return jsonResponse({'error': f'return_format should be one of those formats: {self.RETURN_FORMATS}.'}, status=400)
+        text_manager = TextRequestAdapter(self.oref, versions_params, fill_in_missing_segments, return_format)
+        data = text_manager.get_versions_for_query()
+        data = self._handle_warnings(data)
+        return jsonResponse(data)

build/ci/createJobFromRollout.sh

@@ -0,0 +1,32 @@
+#!/bin/bash
+
+GITHUB_RUN_ID=$1
+DEPLOY_ENV=$2
+
+cat << EOF > job.yaml
+apiVersion: batch/v1
+kind: Job
+metadata:
+  labels:
+    ci-run: "${GITHUB_RUN_ID}"
+    test-name: pytest
+  name: $DEPLOY_ENV-pytest-sandbox-$GITHUB_RUN_ID
+spec:
+  backoffLimit: 2   # in waitForCIJob, we look for 2 fails before declaring failure.  This could be made a variable.
+  template:
+    metadata:
+      labels:
+        ci-run: "${GITHUB_RUN_ID}"
+        test-name: pytest
+    spec:
+EOF
+
+kubectl get rollout $DEPLOY_ENV-web -o yaml | yq '.spec.template.spec' > spec.yaml
+yq -i '.spec.template.spec += load("spec.yaml")' job.yaml
+yq -i '.spec.template.spec.restartPolicy = "Never"' job.yaml
+yq -i '.spec.template.spec.containers[0].args = ["-c", "pip3 install pytest-django; pytest -v -m \"not deep and not failing\" ./sefaria; echo $? > /dev/stdout; exit 0;"]' job.yaml
+yq -i 'del(.spec.template.spec.containers[0].startupProbe)' job.yaml
+yq -i 'del(.spec.template.spec.containers[0].livenessProbe)' job.yaml
+yq -i 'del(.spec.template.spec.containers[0].readinessProbe)' job.yaml
+
+kubectl apply -f job.yaml