Enhance CI

- redone the modelsvalidator
- use the datavalidator from the backend, remove old code
This commit is contained in:
Finn Stutzenstein 2021-08-04 12:44:39 +02:00
parent ec00c6c7db
commit 27d31189c1
No known key found for this signature in database
GPG Key ID: 9042F605C6324654
18 changed files with 421 additions and 1116 deletions

View File

@ -2,34 +2,12 @@
name: Validate models.yml and initial and example data
on: [push, pull_request]
env:
PYTHON_VERSION: 3.8.5
GO_VERSION: 1.16
PYTHON_VERSION: 3.9.6
jobs:
validate-models:
name: Validate models.yml
runs-on: ubuntu-latest
steps:
- name: Set up Go
uses: actions/setup-go@v1
with:
go-version: ${{ env.GO_VERSION }}
- name: Check out code
uses: actions/checkout@v2
- name: Build validator
run: go build
working-directory: docs/modelsvalidator
env:
GO111MODULE: on
- name: Validate models.yml
run: docs/modelsvalidator/modelsvalidator docs/models.yml
validate-data:
name: Validate example-data.json and initial-data.json
runs-on: ubuntu-latest
steps:
- name: Check out code
uses: actions/checkout@v2
@ -39,9 +17,32 @@ jobs:
python-version: ${{ env.PYTHON_VERSION }}
- name: Install requirements
working-directory: docs/datavalidator
run: pip install -U -r requirements.txt
run: pip install -U -r docs/modelsvalidator/requirements.txt
- name: Validate
working-directory: docs/datavalidator
run: python check_json.py
- name: Validate models.yml
working-directory: docs/modelsvalidator
run: python validate.py
validate-data:
name: Validate example-data.json and initial-data.json
runs-on: ubuntu-latest
steps:
- name: Check out code
uses: actions/checkout@v2
- name: Checkout backend submodule
run: git submodule update --init openslides-backend/
- name: Set up Python
uses: actions/setup-python@v2
with:
python-version: ${{ env.PYTHON_VERSION }}
- name: Install requirements
run: pip install -U -r openslides-backend/cli/requirements.txt
- name: set pythonpath
run: echo "PYTHONPATH=openslides-backend" >> $GITHUB_ENV
- name: Validate example-data.json
run: python openslides-backend/cli/check_json.py docs/example-data.json docker/initial-data.json

View File

@ -1,668 +0,0 @@
import json
import re
import sys
from collections import defaultdict
from typing import Any, Callable, Dict, List, Optional, Set, Tuple
import fastjsonschema
import yaml
MODELS_YML_PATH = "../../docs/models.yml"
DEFAULT_FILES = [
"../../docker/initial-data.json",
"../../docs/example-data.json",
]
SCHEMA = fastjsonschema.compile(
{
"$schema": "http://json-schema.org/draft-07/schema#",
"title": "Schema for initial and example data.",
"type": "object",
"patternProperties": {
"^[a-z_]+$": {
"type": "array",
"items": {
"type": "object",
"properties": {"id": {"type": "number"}},
"required": ["id"],
},
}
},
"additionalProperties": False,
}
)
class CheckException(Exception):
pass
def check_string(value: Any) -> bool:
return value is None or isinstance(value, str)
color_regex = re.compile("^#[0-9a-f]{6}$")
def check_color(value: Any) -> bool:
return value is None or bool(isinstance(value, str) and color_regex.match(value))
def check_number(value: Any) -> bool:
return value is None or type(value) == int
def check_float(value: Any) -> bool:
return value is None or type(value) in (int, float)
def check_boolean(value: Any) -> bool:
return value is None or value is False or value is True
def check_string_list(value: Any) -> bool:
return check_x_list(value, check_string)
def check_number_list(value: Any) -> bool:
return check_x_list(value, check_number)
def check_x_list(value: Any, fn: Callable) -> bool:
if value is None:
return True
if not isinstance(value, list):
return False
return all([fn(sv) for sv in value])
def check_decimal(value: Any) -> bool:
if value is None:
return True
if isinstance(value, str):
pattern = r"^-?(\d|[1-9]\d+)\.\d{6}$"
if re.match(pattern, value):
return True
return False
def check_json(value: Any, root: bool = True) -> bool:
if value is None:
return True
if not root and (isinstance(value, int) or isinstance(value, str)):
return True
if isinstance(value, list):
return all(check_json(x, root=False) for x in value)
if isinstance(value, dict):
return all(check_json(x, root=False) for x in value.values())
return False
checker_mapping = {
"string": check_string,
"HTMLStrict": check_string,
"HTMLPermissive": check_string,
"generic-relation": check_string,
"number": check_number,
"timestamp": check_number,
"relation": check_number,
"float": check_float,
"boolean": check_boolean,
"string[]": check_string_list,
"generic-relation-list": check_string_list,
"number[]": check_number_list,
"relation-list": check_number_list,
"decimal(6)": check_decimal,
"color": check_color,
"JSON": check_json,
}
class Checker:
def __init__(self, data: Dict[str, List[Any]], is_import: bool = False) -> None:
self.data = data
with open(MODELS_YML_PATH, "rb") as x:
models_yml = x.read()
models_yml = models_yml.replace(" yes:".encode(), ' "yes":'.encode())
models_yml = models_yml.replace(" no:".encode(), ' "no":'.encode())
self.models = yaml.safe_load(models_yml)
if is_import:
self.modify_models_for_import()
self.errors: List[str] = []
self.template_prefixes: Dict[
str, Dict[str, Tuple[str, int, int]]
] = defaultdict(dict)
self.generate_template_prefixes()
def modify_models_for_import(self) -> None:
collection_allowlist = (
"user",
"meeting",
"group",
"personal_note",
"tag",
"agenda_item",
"list_of_speakers",
"speaker",
"topic",
"motion",
"motion_submitter",
"motion_comment",
"motion_comment_section",
"motion_category",
"motion_block",
"motion_change_recommendation",
"motion_state",
"motion_workflow",
"motion_statute_paragraph",
"poll",
"option",
"vote",
"assignment",
"assignment_candidate",
"mediafile",
"projector",
"projection",
"projector_message",
"projector_countdown",
"chat_group",
)
for collection in list(self.models.keys()):
if collection not in collection_allowlist:
del self.models[collection]
self.models["mediafile"]["blob"] = "string"
def generate_template_prefixes(self) -> None:
for collection in self.models.keys():
for field in self.models[collection]:
if not self.is_template_field(field):
continue
parts = field.split("$")
prefix = parts[0]
suffix = parts[1]
if prefix in self.template_prefixes[collection]:
raise ValueError(
f"the template prefix {prefix} is not unique within {collection}"
)
self.template_prefixes[collection][prefix] = (
field,
len(prefix),
len(suffix),
)
def is_template_field(self, field: str) -> bool:
return "$_" in field or field.endswith("$")
def is_structured_field(self, field: str) -> bool:
return "$" in field and not self.is_template_field(field)
def is_normal_field(self, field: str) -> bool:
return "$" not in field
def is_calculated_field(self, field: str) -> bool:
return field == "content"
def make_structured(self, field: str, replacement: Any) -> str:
if type(replacement) not in (str, int):
raise CheckException(
f"Invalid type {type(replacement)} for the replacement of field {field}"
)
parts = field.split("$")
return parts[0] + "$" + str(replacement) + parts[1]
def to_template_field(
self, collection: str, structured_field: str
) -> Tuple[str, str]:
"""Returns template_field, replacement"""
parts = structured_field.split("$")
descriptor = self.template_prefixes[collection].get(parts[0])
if not descriptor:
raise CheckException(
f"Unknown template field for prefix {parts[0]} in collection {collection}"
)
return (
descriptor[0],
structured_field[descriptor[1] + 1 : len(structured_field) - descriptor[2]],
)
def run_check(self) -> None:
self.check_json()
self.check_collections()
for collection, models in self.data.items():
for model in models:
self.check_model(collection, model)
if self.errors:
errors = [f"\t{error}" for error in self.errors]
raise CheckException("\n".join(errors))
def check_json(self) -> None:
try:
SCHEMA(self.data)
except fastjsonschema.exceptions.JsonSchemaException as e:
raise CheckException(f"JSON does not match schema: {str(e)}")
def check_collections(self) -> None:
c1 = set(self.data.keys())
c2 = set(self.models.keys())
if c1 != c2:
err = "Collections in JSON file do not match with models.yml."
if c2 - c1:
err += f" Missing collections: {', '.join(c2-c1)}."
if c1 - c2:
err += f" Invalid collections: {', '.join(c1-c2)}."
raise CheckException(err)
def check_model(self, collection: str, model: Dict[str, Any]) -> None:
errors = self.check_normal_fields(model, collection)
if not errors:
errors = self.check_template_fields(model, collection)
if not errors:
self.check_types(model, collection)
self.check_relations(model, collection)
def check_normal_fields(self, model: Dict[str, Any], collection: str) -> bool:
model_fields = set(
x
for x in model.keys()
if self.is_normal_field(x) or self.is_template_field(x)
)
collection_fields = set(
x
for x in self.models[collection].keys()
if self.is_normal_field(x) or self.is_template_field(x)
)
calculated_fields = set(
x
for x in self.models[collection].keys()
if self.is_calculated_field(x)
)
errors = False
if collection_fields - model_fields - calculated_fields:
error = f"{collection}/{model['id']}: Missing fields {', '.join(collection_fields - model_fields)}"
self.errors.append(error)
errors = True
if model_fields - collection_fields:
error = f"{collection}/{model['id']}: Invalid fields {', '.join(model_fields - collection_fields)}"
self.errors.append(error)
errors = True
return errors
def check_template_fields(self, model: Dict[str, Any], collection: str) -> bool:
"""
Only checks that for each replacement a structured field exists and
not too many structured fields. Does not check the content.
Returns True on errors.
"""
errors = False
for template_field in self.models[collection].keys():
if not self.is_template_field(template_field):
continue
field_error = False
replacements = model[template_field]
if not isinstance(replacements, list):
self.errors.append(
f"{collection}/{model['id']}/{template_field}: Replacements for the template field must be a list"
)
field_error = True
continue
for replacement in replacements:
if not isinstance(replacement, str):
self.errors.append(
f"{collection}/{model['id']}/{template_field}: Each replacement for the template field must be a string"
)
field_error = True
if field_error:
error = True
continue
replacement_collection = None
field_description = self.models[collection][template_field]
if isinstance(field_description, dict):
replacement_collection = field_description.get("replacement_collection")
for replacement in replacements:
structured_field = self.make_structured(template_field, replacement)
if structured_field not in model:
self.errors.append(
f"{collection}/{model['id']}/{template_field}: Missing {structured_field} since it is given as a replacement"
)
errors = True
if replacement_collection:
try:
as_id = int(replacement)
except (TypeError, ValueError):
self.errors.append(
f"{collection}/{model['id']}/{template_field}: Replacement {replacement} is not an integer"
)
if not self.find_model(replacement_collection, as_id):
self.errors.append(
f"{collection}/{model['id']}/{template_field}: Replacement {replacement} does not exist as a model of collection {replacement_collection}"
)
for field in model.keys():
if self.is_structured_field(field):
try:
_template_field, _replacement = self.to_template_field(
collection, field
)
if (
template_field == _template_field
and _replacement not in model[template_field]
):
self.errors.append(
f"{collection}/{model['id']}/{field}: Invalid structured field. Missing replacement {_replacement} in {template_field}"
)
errors = True
except CheckException as e:
self.errors.append(
f"{collection}/{model['id']}/{field} error: " + str(e)
)
errors = True
return errors
def check_types(self, model: Dict[str, Any], collection: str) -> None:
for field in model.keys():
if self.is_template_field(field):
continue
field_type = self.get_type_from_collection(field, collection)
enum = self.get_enum_from_collection_field(field, collection)
checker = checker_mapping.get(field_type)
if checker is None:
raise NotImplementedError(
f"TODO implement check for field type {field_type}"
)
if not checker(model[field]):
error = f"{collection}/{model['id']}/{field}: Type error: Type is not {field_type}"
self.errors.append(error)
if enum and model[field] not in enum:
error = f"{collection}/{model['id']}/{field}: Value error: Value {model[field]} is not a valid enum value"
self.errors.append(error)
def get_type_from_collection(self, field: str, collection: str) -> str:
if self.is_structured_field(field):
field, _ = self.to_template_field(collection, field)
field_description = self.models[collection][field]
if isinstance(field_description, dict):
if field_description["type"] == "template":
if isinstance(field_description["fields"], dict):
return field_description["fields"]["type"]
return field_description["fields"]
return field_description["type"]
return field_description
field_description = self.get_field_description(field, collection)
if field_description:
return field_description["type"]
return self.models[collection][field]
def get_enum_from_collection_field(
self, field: str, collection: str
) -> Optional[Set[str]]:
if self.is_structured_field(field):
field, _ = self.to_template_field(collection, field)
field_description = self.models[collection][field]
if isinstance(field_description, dict):
if field_description["type"] == "template":
if isinstance(field_description["fields"], dict):
field_description = field_description["fields"]
else:
return None
if "enum" in field_description:
enum = set(field_description["enum"])
if not field_description.get("required", False):
enum.add(None)
return enum
return None
def check_relations(self, model: Dict[str, Any], collection: str) -> None:
for field in model.keys():
try:
self.check_relation(model, collection, field)
except CheckException as e:
self.errors.append(
f"{collection}/{model['id']}/{field} error: " + str(e)
)
def check_relation(
self, model: Dict[str, Any], collection: str, field: str
) -> None:
if self.is_template_field(field):
return
field_type = self.get_type_from_collection(field, collection)
basemsg = f"{collection}/{model['id']}/{field}: Relation Error: "
replacement = None
if self.is_structured_field(field):
_, replacement = self.to_template_field(collection, field)
if field_type == "relation":
foreign_id = model[field]
if foreign_id is None:
return
foreign_collection, foreign_field = self.get_to(field, collection)
self.check_reverse_relation(
collection,
model["id"],
model,
foreign_collection,
foreign_id,
foreign_field,
basemsg,
replacement,
)
elif field_type == "relation-list":
foreign_ids = model[field]
if foreign_ids is None:
return
foreign_collection, foreign_field = self.get_to(field, collection)
for foreign_id in foreign_ids:
self.check_reverse_relation(
collection,
model["id"],
model,
foreign_collection,
foreign_id,
foreign_field,
basemsg,
replacement,
)
elif field_type == "generic-relation" and model[field] is not None:
foreign_collection, foreign_id = self.split_fqid(model[field])
foreign_field = self.get_to_generic_case(
collection, field, foreign_collection
)
self.check_reverse_relation(
collection,
model["id"],
model,
foreign_collection,
foreign_id,
foreign_field,
basemsg,
replacement,
)
elif field_type == "generic-relation-list" and model[field] is not None:
for fqid in model[field]:
foreign_collection, foreign_id = self.split_fqid(fqid)
foreign_field = self.get_to_generic_case(
collection, field, foreign_collection
)
self.check_reverse_relation(
collection,
model["id"],
model,
foreign_collection,
foreign_id,
foreign_field,
basemsg,
replacement,
)
def get_to(self, field: str, collection: str) -> Tuple[str, str]:
if self.is_structured_field(field):
field, _ = self.to_template_field(collection, field)
field_description = self.models[collection][field]
if field_description["type"] == "template":
to = field_description["fields"]["to"]
else:
to = field_description["to"]
return to.split("/")
def find_model(self, collection: str, id: int) -> Optional[Dict[str, Any]]:
c = self.data.get(collection, [])
for model in c:
if model["id"] == id:
return model
return None
def check_reverse_relation(
self,
collection: str,
id: int,
model: Dict[str, Any],
foreign_collection: str,
foreign_id: int,
foreign_field: str,
basemsg: str,
replacement: Optional[str],
) -> None:
actual_foreign_field = foreign_field
if self.is_template_field(foreign_field):
if replacement:
actual_foreign_field = self.make_structured(foreign_field, replacement)
else:
replacement_collection = self.models[foreign_collection][foreign_field][
"replacement_collection"
]
replacement = model.get(f"{replacement_collection}_id")
if not replacement:
self.errors.append(
f"{basemsg} points to {foreign_collection}/{foreign_id}/{foreign_field},"
f" but there is no replacement for {replacement_collection}"
)
actual_foreign_field = self.make_structured(foreign_field, replacement)
foreign_model = self.find_model(foreign_collection, foreign_id)
foreign_value = (
foreign_model.get(actual_foreign_field)
if foreign_model is not None
else None
)
foreign_field_type = self.get_type_from_collection(
foreign_field, foreign_collection
)
fqid = f"{collection}/{id}"
error = False
if foreign_field_type == "relation":
error = foreign_value != id
elif foreign_field_type == "relation-list":
error = not foreign_value or id not in foreign_value
elif foreign_field_type == "generic-relation":
error = foreign_value != fqid
elif foreign_field_type == "generic-relation-list":
error = not foreign_value or fqid not in foreign_value
else:
raise NotImplementedError()
if error:
self.errors.append(
f"{basemsg} points to {foreign_collection}/{foreign_id}/{actual_foreign_field},"
" but the reverse relation for is corrupt"
)
def split_fqid(self, fqid: str) -> Tuple[str, int]:
try:
collection, _id = fqid.split("/")
id = int(_id)
if collection not in self.models.keys():
raise CheckException(f"Fqid {fqid} has an invalid collection")
return collection, id
except (ValueError, AttributeError):
raise CheckException(f"Fqid {fqid} is malformed")
def split_collectionfield(self, collectionfield: str) -> Tuple[str, str]:
collection, field = collectionfield.split("/")
if collection not in self.models.keys():
raise CheckException(
f"Collectionfield {collectionfield} has an invalid collection"
)
if (
field not in self.models[collection]
): # Note: this has to be adopted when supporting template fields
raise CheckException(
f"Collectionfield {collectionfield} has an invalid field"
)
return collection, field
def get_to_generic_case(
self, collection: str, field: str, foreign_collection: str
) -> str:
"""Returns all reverse relations as collectionfields"""
to = self.models[collection][field]["to"]
if isinstance(to, dict):
if foreign_collection not in to["collections"]:
raise CheckException(
f"The collection {foreign_collection} is not supported "
"as a reverse relation in {collection}/{field}"
)
return to["field"]
for cf in to:
c, f = self.split_collectionfield(cf)
if c == foreign_collection:
return f
raise CheckException(
f"The collection {foreign_collection} is not supported as a reverse relation in {collection}/{field}"
)
def main() -> int:
files = sys.argv[1:]
if not files:
files = DEFAULT_FILES
is_import = "--import" in files
if is_import:
files = [x for x in files if x != "--import"]
failed = False
for f in files:
with open(f) as data:
try:
Checker(json.load(data), is_import=is_import).run_check()
except CheckException as e:
print(f"Check for {f} failed:\n", e)
failed = True
else:
print(f"Check for {f} successful.")
return 1 if failed else 0
if __name__ == "__main__":
sys.exit(main())

File diff suppressed because one or more lines are too long

View File

@ -1,2 +0,0 @@
fastjsonschema
pyyaml

View File

@ -52,7 +52,7 @@
# for all the fields that come from the template field.
# JSON Schema Properties:
# - You can add JSON Schema properties to the fields like `enum`, `description`,
# `maxLength` and `minimum`
# `items`, `maxLength` and `minimum`
# Additional properties:
# - The property `read_only` describes a field that can not be changed by an action.
# - The property `default` describes the default value that is used for new objects.
@ -60,6 +60,9 @@
# string. If this field is given it must have some content.
# - The property `equal_fields` describes fields that must have the same value in
# the instance and the related instance.
# Restriction Mode:
# The field `restriction_mode` is required for every field. It puts the field into a
# restriction group. See https://github.com/OpenSlides/OpenSlides/wiki/Restrictions-Overview
organization:
id:
@ -463,6 +466,7 @@ meeting:
restriction_mode: B
imported_at:
type: timestamp
restriction_mode: B
# Configuration (only for the server owner)
jitsi_domain:
@ -2393,10 +2397,10 @@ option:
text:
type: HTMLStrict
restriction_mode: A
yes:
"yes":
type: decimal(6)
restriction_mode: B
no:
"no":
type: decimal(6)
restriction_mode: B
abstain:
@ -2589,7 +2593,7 @@ mediafile:
restriction_mode: B
filename:
type: string
descriptin: The uploaded filename. Will be used for downloading. Only writeable on create.
description: The uploaded filename. Will be used for downloading. Only writeable on create.
restriction_mode: B
mimetype:
type: string
@ -2797,6 +2801,11 @@ projection:
type: string
restriction_mode: A
content:
type: JSON
calculated: true
restriction_mode: A
current_projector_id:
type: relation
to: projector/current_projection_ids
@ -2834,10 +2843,6 @@ projection:
to: meeting/all_projection_ids
required: true
restriction_mode: A
content:
type: JSON
calculated: true
restriction_mode: A
projector_message:
id:

View File

@ -1,21 +0,0 @@
# Modelsvalidator
Modelsvalidator is a tool to validate the models.yml file
## Run
Build first: `go build ./...`.
The tool requires the content of the models.yml. It can be provided via stdin, a
file system path or an url starting with http:// or https://.
```
cat models.yml | modelsvalidator
modelsvalidator openslides/docs/models.yml
modelsvalidator https://raw.githubusercontent.com/OpenSlides/OpenSlides/openslides4-dev/docs/models.yml
```
The tool returns with status code 0 and no content, if the given content is
valid. It returns with a positive status code and some error messages if not.

View File

@ -1,159 +0,0 @@
package check
import (
"fmt"
"strings"
models "github.com/OpenSlides/openslides-models-to-go"
)
// Check runs some checks on the given models.
func Check(data map[string]models.Model) error {
validators := []func(map[string]models.Model) error{
validateTypes,
validateRelations,
validateTemplatePrefixes,
}
errors := new(ErrorList)
for _, v := range validators {
if err := v(data); err != nil {
errors.append(err)
}
}
if !errors.empty() {
return errors
}
return nil
}
func validateTypes(data map[string]models.Model) error {
scalar := scalarTypes()
relation := relationTypes()
errs := &ErrorList{
Name: "type validator",
intent: 1,
}
for modelName, model := range data {
for fieldName, field := range model.Fields {
if scalar[strings.TrimSuffix(field.Type, "[]")] {
continue
}
if relation[field.Type] {
continue
}
errs.append(fmt.Errorf("Unknown type `%s` in %s/%s", field.Type, modelName, fieldName))
}
}
if errs.empty() {
return nil
}
return errs
}
func validateRelations(data map[string]models.Model) error {
errs := &ErrorList{
Name: "relation validator",
intent: 1,
}
relation := relationTypes()
for modelName, model := range data {
Next:
for fieldName, field := range model.Fields {
r := field.Relation()
if r == nil {
continue
}
for _, c := range r.ToCollections() {
toModel, ok := data[c.Collection]
if !ok {
errs.append(fmt.Errorf("%s/%s directs to nonexisting model `%s`", modelName, fieldName, c.Collection))
continue Next
}
toField, ok := toModel.Fields[c.ToField.Name]
if !ok {
errs.append(fmt.Errorf("%s/%s directs to nonexisting collectionfield `%s/%s`", modelName, fieldName, c.Collection, c.ToField.Name))
continue Next
}
if !relation[toField.Type] {
errs.append(fmt.Errorf("%s/%s directs to `%s/%s`, but it is not a relation, but %s", modelName, fieldName, c.Collection, c.ToField.Name, toField.Type))
continue Next
}
}
}
}
if errs.empty() {
return nil
}
return errs
}
func validateTemplatePrefixes(models map[string]models.Model) error {
errs := &ErrorList{
Name: "template prefixes validator",
intent: 1,
}
for modelName, model := range models {
prefixes := map[string]bool{}
for fieldName := range model.Fields {
i := strings.Index(fieldName, "$")
if i < 0 {
continue
}
prefix := fieldName[0:i]
if prefixes[prefix] {
errs.append(fmt.Errorf("Duplicate template prefix %s in %s", prefix, modelName))
}
prefixes[prefix] = true
}
}
if errs.empty() {
return nil
}
return errs
}
// scalarTypes are the main types. All scalarTypes can be used as a list.
// JSON[], timestamp[] etc.
func scalarTypes() map[string]bool {
s := []string{
"string",
"number",
"boolean",
"JSON",
"HTMLPermissive",
"HTMLStrict",
"float",
"decimal(6)",
"timestamp",
"color",
}
out := make(map[string]bool)
for _, t := range s {
out[t] = true
}
return out
}
// relationTypes are realtion types in realtion to other fields.
func relationTypes() map[string]bool {
s := []string{
"relation",
"relation-list",
"generic-relation",
"generic-relation-list",
"template",
}
out := make(map[string]bool)
for _, t := range s {
out[t] = true
}
return out
}

View File

@ -1,121 +0,0 @@
package check_test
import (
"errors"
"strings"
"testing"
"github.com/OpenSlides/Openslides/modelsvalidator/check"
models "github.com/OpenSlides/openslides-models-to-go"
)
const yamlUnknownFieldType = `---
some_model:
field: unknown
`
const yamlNonExistingModel = `---
some_model:
no_other_model:
type: relation
to: not_existing/field
`
const yamlNonExistingField = `---
some_model:
no_other_field:
type: relation
to: other_model/bar
other_model:
foo: string
`
const yamlDuplicateTemplatePrefix = `---
some_model:
field_$_1: number
field_$_2: number
`
const yamlWrongReverseRelaitonType = `---
some_model:
other_model:
type: relation
to: other_model/field
other_model:
field: HTMLStrict
`
func TestCheck(t *testing.T) {
for _, tt := range []struct {
name string
yaml string
err string
}{
{
"unknown type",
yamlUnknownFieldType,
"Unknown type `unknown` in some_model/field",
},
{
"non-existing model",
yamlNonExistingModel,
"some_model/no_other_model directs to nonexisting model `not_existing`",
},
{
"non-existing Field",
yamlNonExistingField,
"some_model/no_other_field directs to nonexisting collectionfield `other_model/bar`",
},
{
"duplicate template prefix",
yamlDuplicateTemplatePrefix,
"Duplicate template prefix field_ in some_model",
},
{
"wrong reverse relation type",
yamlWrongReverseRelaitonType,
"some_model/other_model directs to `other_model/field`, but it is not a relation, but HTMLStrict",
},
} {
t.Run(tt.name, func(t *testing.T) {
data, err := models.Unmarshal(strings.NewReader(tt.yaml))
if err != nil {
t.Fatalf("Can not unmarshal yaml: %v", err)
}
gotErr := check.Check(data)
if tt.err == "" {
if gotErr != nil {
t.Errorf("Models.Check() returned an unexepcted error: %v", err)
}
return
}
if gotErr == nil {
t.Fatalf("Models.Check() did not return an error, expected: %v", tt.err)
}
var errList *check.ErrorList
if !errors.As(gotErr, &errList) {
t.Fatalf("Models.Check() did not return a ListError, got: %v", gotErr)
}
var found bool
for _, err := range errList.Errs {
var errList *check.ErrorList
if !errors.As(err, &errList) {
continue
}
for _, err := range errList.Errs {
if err.Error() == tt.err {
found = true
}
}
}
if !found {
t.Errorf("Models.Check() returned:\n%v\n\nExpected something that contains:\n%v", gotErr, tt.err)
}
})
}
}

View File

@ -1,38 +0,0 @@
package check
import (
"fmt"
"strings"
)
// ErrorList is an error that contains a list of other errors.
type ErrorList struct {
Name string
intent int
Errs []error
}
func (e *ErrorList) append(err error) {
if err == nil {
return
}
e.Errs = append(e.Errs, err)
}
func (e ErrorList) Error() string {
intent := strings.Repeat(" ", e.intent)
var msgs []string
for _, err := range e.Errs {
msgs = append(msgs, fmt.Sprintf("%s* %v", intent, err))
}
msg := strings.Join(msgs, "\n")
if e.Name != "" {
return fmt.Sprintf("%s:\n%s", e.Name, msg)
}
return msg
}
func (e *ErrorList) empty() bool {
return len(e.Errs) == 0
}

View File

@ -1,5 +0,0 @@
module github.com/OpenSlides/Openslides/modelsvalidator
go 1.15
require github.com/OpenSlides/openslides-models-to-go v0.2.0

View File

@ -1,6 +0,0 @@
github.com/OpenSlides/openslides-models-to-go v0.2.0 h1:1D+rD94GcUZAvmHLRyzJCIj7wklU+JBkoYifwtHH08s=
github.com/OpenSlides/openslides-models-to-go v0.2.0/go.mod h1:CriCefW5smTixhFfVLiuA8NgyMX4PAU5e3YpJHnaZx8=
gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405 h1:yhCVgyC4o1eVCa2tZl7eS0r+SDo693bJlVdllGtEeKM=
gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
gopkg.in/yaml.v3 v3.0.0-20200615113413-eeeca48fe776 h1:tQIYjPdBoyREyB9XMu+nnTclpTYkz2zFM+lzLJFO4gQ=
gopkg.in/yaml.v3 v3.0.0-20200615113413-eeeca48fe776/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=

View File

@ -1,56 +0,0 @@
package main
import (
"fmt"
"io"
"log"
"net/http"
"os"
"strings"
"github.com/OpenSlides/Openslides/modelsvalidator/check"
models "github.com/OpenSlides/openslides-models-to-go"
)
func main() {
var content io.Reader = os.Stdin
if len(os.Args) > 1 {
c, err := openModels(os.Args[1])
if err != nil {
log.Fatalf("Can not load content: %v", err)
}
defer c.Close()
content = c
}
data, err := models.Unmarshal(content)
if err != nil {
log.Fatalf("Invalid model format: %v", err)
}
if err := check.Check(data); err != nil {
log.Fatalf("Invalid model structure:\n\n%v", err)
}
}
// openModels reads the model either from file or from an url.
func openModels(path string) (io.ReadCloser, error) {
if strings.HasPrefix(path, "http://") || strings.HasPrefix(path, "https://") {
return openModelsFromURL(path)
}
return os.Open(path)
}
func openModelsFromURL(url string) (io.ReadCloser, error) {
resp, err := http.Get(url)
if err != nil {
return nil, fmt.Errorf("requesting models from url: %w", err)
}
if resp.StatusCode != 200 {
return nil, fmt.Errorf("can not get models from url. Got status %s", resp.Status)
}
return resp.Body, nil
}

View File

@ -0,0 +1 @@
pyyaml

View File

@ -0,0 +1,375 @@
import json
import re
import sys
from collections import defaultdict
from typing import Any, Callable, Dict, List, Optional, Set, Tuple, Union
import yaml
DEFAULT_FILES = [
"../../docs/models.yml",
]
KEYSEPARATOR = "/"
_collection_regex = r"[a-z]([a-z_]+[a-z]+)?"
_field_regex = r"[a-z][a-z0-9_]*\$?[a-z0-9_]*"
collectionfield_regex = re.compile(
f"^({_collection_regex}){KEYSEPARATOR}({_field_regex})$"
)
collection_regex = re.compile(f"^{_collection_regex}$")
field_regex = re.compile(f"^{_field_regex}$")
decimal_regex = re.compile("^\d+\.\d{6}$")
color_regex = re.compile("^#[0-9a-f]{6}$")
RELATION_TYPES = (
"relation",
"relation-list",
"generic-relation",
"generic-relation-list",
)
DATA_TYPES = (
"string",
"number",
"string[]",
"number[]",
"boolean",
"JSON",
"HTMLStrict",
"HTMLPermissive",
"float",
"decimal(6)",
"timestamp",
"color",
)
VALID_TYPES = DATA_TYPES + RELATION_TYPES + ("template",)
OPTIONAL_ATTRIBUTES = (
"description",
"calculated",
"required",
"read_only",
)
class CheckException(Exception):
pass
class Checker:
def __init__(self, filepath: str) -> None:
with open(filepath, "rb") as x:
self.models = yaml.safe_load(x.read())
self.errors: List[str] = []
def run_check(self) -> None:
self._run_checks()
if self.errors:
errors = [f"\t{error}" for error in self.errors]
raise CheckException("\n".join(errors))
def _run_checks(self) -> None:
for collection in self.models.keys():
if not collection_regex.match(collection):
self.errors.append(f"Collection '{collection}' is not valid.")
if self.errors:
return
for collection, fields in self.models.items():
if not isinstance(fields, dict):
self.errors.append(
f"The fields of collection {collection} must be a dict."
)
continue
for field_name, field in fields.items():
if not field_regex.match(field_name):
self.errors.append(
f"Field name '{field_name}' of collection {collection} is not a valid field name."
)
continue
if not isinstance(field, dict):
self.errors.append(
f"Field '{field_name}' of collection {collection} must be a dict."
)
self.check_field(collection, field_name, field)
if self.errors:
return
for collection, fields in self.models.items():
for field_name, field in fields.items():
is_relation_field = field["type"] in RELATION_TYPES
is_template_relation_field = (
field["type"] == "template"
and isinstance(field["fields"], dict)
and field["fields"]["type"] in RELATION_TYPES
)
if not is_relation_field and not is_template_relation_field:
continue
error = self.check_relation(collection, field_name, field)
if error:
self.errors.append(error)
def check_field(
self,
collection: str,
field_name: str,
field: Union[str, Dict[str, Any]],
nested: bool = False,
) -> None:
collectionfield = f"{collection}{KEYSEPARATOR}{field_name}"
if nested:
if isinstance(field, str):
field = {"type": field}
field[
"restriction_mode"
] = "A" # add restriction_mode to satisfy the checker below.
if field["type"] == "template": # no nested templates
self.errors.append(f"Nested template field in {collectionfield}")
return
type = field.get("type")
if type not in VALID_TYPES:
self.errors.append(
f"Type '{type}' for collectionfield {collectionfield} is invalid."
)
return
required_attributes = [
"type",
"restriction_mode",
]
if type in RELATION_TYPES:
required_attributes.append("to")
if type == "template":
required_attributes.append("fields")
for attr in required_attributes:
if attr not in field:
self.errors.append(
f"Required attribute '{attr}' for collectionfield {collectionfield} is missing."
)
return
if field.get("calculated"):
return
valid_attributes = list(OPTIONAL_ATTRIBUTES) + required_attributes
if type == "string[]":
valid_attributes.append("items")
if "items" in field and "enum" not in field["items"]:
self.errors.append(
f"'items' is missing an inner 'enum' for {collectionfield}"
)
return
for value in field.get("items", {"enum": []})["enum"]:
self.validate_value_for_type("string", value, collectionfield)
if type == "JSON" and "default" in field:
try:
json.loads(json.dumps(field["default"]))
except:
self.errors.append(
f"Default value for {collectionfield}' is not valid json."
)
if type == "number":
valid_attributes.append("minimum")
if not isinstance(field.get("minimum", 0), int):
self.errors.append(f"'minimum' for {collectionfield} is not a number.")
if type == "string":
valid_attributes.append("maxLength")
if not isinstance(field.get("maxLength", 0), int):
self.errors.append(
f"'maxLength' for {collectionfield} is not a number."
)
if type in DATA_TYPES:
valid_attributes.append("default")
if "default" in field:
self.validate_value_for_type(type, field["default"], collectionfield)
valid_attributes.append("enum")
if "enum" in field:
if not isinstance(field["enum"], list):
self.errors.append(f"'enum' for {collectionfield}' is not a list.")
for value in field["enum"]:
self.validate_value_for_type(type, value, collectionfield)
if type in RELATION_TYPES:
valid_attributes.append("on_delete")
if "on_delete" in field and field["on_delete"] not in (
"CASCADE",
"PROTECT",
):
self.errors.append(
f"invalid value for 'on_delete' for {collectionfield}"
)
valid_attributes.append("equal_fields")
if type == "template":
if "$" not in field_name:
self.errors.append(
f"The template field {collectionfield} is missing a $"
)
valid_attributes.append("replacement_collection")
elif "$" in field_name and not nested:
print(field_name, field)
self.errors.append(f"The non-template field {collectionfield} contains a $")
for attr in field.keys():
if attr not in valid_attributes:
self.errors.append(
f"Attribute '{attr}' for collectionfield {collectionfield} is invalid."
)
if not isinstance(field.get("description", ""), str):
self.errors.append(f"Description of {collectionfield} must be a string.")
if type == "template":
self.check_field(collection, field_name, field["fields"], nested=True)
def validate_value_for_type(
self, type_str: str, value: Any, collectionfield: str
) -> None:
basic_types = {
"string": str,
"number": int,
"boolean": bool,
"HTMLStrict": str,
"HTMLPermissive": str,
"timestamp": int,
}
if type_str in basic_types:
if type(value) != basic_types[type_str]:
self.errors.append(
f"Value '{value}' for {collectionfield}' is not a {type_str}."
)
elif type_str in ("string[]", "number[]"):
if not isinstance(value, list):
self.errors.append(
f"Value '{value}' for {collectionfield}' is not a {type_str}."
)
for x in value:
if type(x) != basic_types[type_str[:-2]]:
self.errors.append(
f"Listentry '{x}' for {collectionfield}' is not a {type_str[:-2]}."
)
elif type_str == "JSON":
pass
elif type_str == "float":
if type(value) not in (int, float):
self.errors.append(
f"Value '{value}' for {collectionfield}' is not a float."
)
elif type_str == "decimal(6)":
if not decimal_regex.match(value):
self.errors.append(
f"Value '{value}' for {collectionfield}' is not a decimal(6)."
)
elif type_str == "color":
if not color_regex.match(value):
self.errors.append(
f"Value '{value}' for {collectionfield}' is not a color."
)
else:
raise NotImplementedError(type_str)
def check_relation(
self, collection: str, field_name: str, field: Dict[str, Any]
) -> Optional[str]:
collectionfield = f"{collection}{KEYSEPARATOR}{field_name}"
if field["type"] == "template":
field = field["fields"]
to = field["to"]
if isinstance(to, str):
if not collectionfield_regex.match(to):
return f"'to' of {collectionfield} is not a collectionfield."
return self.check_reverse(collectionfield, to)
elif isinstance(to, list):
for cf in to:
if not collectionfield_regex.match(cf):
return f"The collectionfield in 'to' of {collectionfield} is not valid."
error = self.check_reverse(collectionfield, cf)
if error:
return error
else:
to_field = to["field"]
if not field_regex.match(to_field):
return (
f"The field '{to_field}' in 'to' of {collectionfield} is not valid."
)
for c in to["collections"]:
if not collection_regex.match(c):
self.errors.append(
f"The collection '{c}' in 'to' of {collectionfield} is not a valid collection."
)
error = self.check_reverse(
collectionfield, f"{c}{KEYSEPARATOR}{to['field']}"
)
if error:
return error
return None
def check_reverse(
self, from_collectionfield: str, to_collectionfield: str
) -> Optional[str]:
to_unified = [] # a list of target collectionfields (unififed with all
# the different possibilities for the 'to' field) from the (expected)
# relation in to_collectionfield. The from_collectionfield must be in this
# list.
to_collection, to_field_name = to_collectionfield.split(KEYSEPARATOR)
if to_collection not in self.models:
return f"The collection '{to_collection}' in 'to' of {from_collectionfield} is not a valid collection."
if to_field_name not in self.models[to_collection]:
return f"The collectionfield '{to_collectionfield}' in 'to' of {from_collectionfield} does not exist."
to_field = self.models[to_collection][to_field_name]
if to_field["type"] == "template":
to_field = to_field["fields"]
if not isinstance(to_field, dict):
return f"The 'fields' of the template field '{to_collectionfield}' must be a dict to hold a relation."
if to_field["type"] not in RELATION_TYPES:
return f"{from_collectionfield} points to {to_collectionfield}, but {to_collectionfield} to is not a relation."
to = to_field["to"]
if isinstance(to, str):
to_unified.append(to)
elif isinstance(to, list):
to_unified = to
else:
for c in to["collections"]:
to_unified.append(f"{c}{KEYSEPARATOR}{to['field']}")
if from_collectionfield not in to_unified:
return f"{from_collectionfield} points to {to_collectionfield}, but {to_collectionfield} does not point back."
return None
def split_collectionfield(self, collectionfield: str) -> Tuple[str, str]:
parts = collectionfield.split(KEYSEPARATOR)
return parts[0], parts[1]
def main() -> int:
files = sys.argv[1:]
if not files:
files = DEFAULT_FILES
failed = False
for f in files:
with open(f) as data:
try:
Checker(f).run_check()
except CheckException as e:
print(f"Check for {f} failed:\n", e)
failed = True
else:
print(f"Check for {f} successful.")
return 1 if failed else 0
if __name__ == "__main__":
sys.exit(main())

@ -1 +1 @@
Subproject commit 4bf228dd0e246a466c403d31bd0158242c756f96
Subproject commit 55fa585ff740c1bac9511228ec55a331997432ab

@ -1 +1 @@
Subproject commit eb1e7fd9ce2463c1a40a33b1fc65fbcc503035d1
Subproject commit 6f39f6074da8f4936f3127652097a4d1bafb4e32

@ -1 +1 @@
Subproject commit ad603239769e00345aa7855cb45ac71dfb08e396
Subproject commit ce2b08ef5fc0d677a6a4f20c0ed9ec120181cadf