Skip to content

Commit

Permalink
docs/policy-language: reorder Universal Quantification content (open-…
Browse files Browse the repository at this point in the history
…policy-agent#4605)

We should really start with `every`, then offer alternatives, and
call out common mistakes in those alternatives after having presented
the proper solution.

Also replaces some one-letter-rule names in the detailled presentation
of `every`.

Fixes open-policy-agent#4603.

Signed-off-by: Stephan Renatus <stephan.renatus@gmail.com>
  • Loading branch information
srenatus committed Apr 22, 2022
1 parent 1890b0e commit 1ca73a0
Showing 1 changed file with 67 additions and 55 deletions.
122 changes: 67 additions & 55 deletions docs/content/policy-language.md
Expand Up @@ -1031,53 +1031,40 @@ apps_not_in_prod[name]

## Universal Quantification (FOR ALL)

Like SQL, Rego does not have a direct way to express _universal quantification_
("FOR ALL"). However, like SQL, you can use other language primitives (e.g.,
[Negation](#negation)) to express FOR ALL. For example, imagine you want to
express a policy that says (in English):
Rego allows for several ways to express universal quantification.

For example, imagine you want to express a policy that says (in English):

```
There must be no apps named "bitcoin-miner".
```

A common mistake is to try encoding the policy with a rule named
`no_bitcoin_miners` like so:
The most expressive way to state this in Rego is using the `every` keyword:

```live:eg/data/incorrect_no_bitcoin:module:read_only
no_bitcoin_miners {
app := apps[_]
app.name != "bitcoin-miner" # THIS IS NOT CORRECT.
}
```

It becomes clear that this is incorrect when you use the [`some`](#some-keyword)
keyword, because the rule is true whenever there is SOME app that is not a
bitcoin-miner:

```live:eg/data/incorrect_no_bitcoin_some:module
import future.keywords.in
```live:eg/data/every_alternative:module:read_only
import future.keywords.every
no_bitcoin_miners {
some app in apps
app.name != "bitcoin-miner"
no_bitcoin_miners_using_every {
every app in apps {
app.name != "bitcoin-miner"
}
}
```

You can confirm this by querying the rule:
Variables in Rego are _existentially quantified_ by default: when you write

```live:eg/data/incorrect_no_bitcoin_some:query:merge_down
no_bitcoin_miners with apps as [{"name": "bitcoin-miner"}, {"name": "web"}]
```live:eg/data/every_alternative/1:query:merge_down
array := ["one", "two", "three"]; array[i] == "three"
```
```live:eg/data/incorrect_no_bitcoin_some:output

The query will be satisfied **if there is an `i`** such that the query's
expressions are simultaneously satisfied.
```live:eg/data/every_alternative/1:output
```

The reason the rule is incorrect is that variables in Rego are _existentially
quantified_. This means that rule bodies and queries express FOR ANY and not FOR
ALL. To express FOR ALL in Rego complement the logic in the rule body (e.g.,
`!=` becomes `==`) and then complement the check using negation (e.g.,
`no_bitcoin_miners` becomes `not any_bitcoin_miners`).
Therefore, there are other ways to express the desired policy.

For this policy, you define a rule that finds if there exists a bitcoin-mining
For this policy, you can also define a rule that finds if there exists a bitcoin-mining
app (which is easy using the `some` keyword). And then you use negation to check
that there is NO bitcoin-mining app. Technically, you're using 2 negations and
an existential quantifier, which is logically the same as a universal
Expand Down Expand Up @@ -1116,6 +1103,43 @@ value for `no_bitcoin_miners_using_negation`. Since the body of the rule fails
to match, there is no value generated.
{{< /info >}}

A common mistake is to try encoding the policy with a rule named `no_bitcoin_miners`
like so:

```live:eg/data/incorrect_no_bitcoin:module:read_only
no_bitcoin_miners {
app := apps[_]
app.name != "bitcoin-miner" # THIS IS NOT CORRECT.
}
```

It becomes clear that this is incorrect when you use the [`some`](#some-keyword)
keyword, because the rule is true whenever there is SOME app that is not a
bitcoin-miner:

```live:eg/data/incorrect_no_bitcoin_some:module
import future.keywords.in
no_bitcoin_miners {
some app in apps
app.name != "bitcoin-miner"
}
```

You can confirm this by querying the rule:

```live:eg/data/incorrect_no_bitcoin_some:query:merge_down
no_bitcoin_miners with apps as [{"name": "bitcoin-miner"}, {"name": "web"}]
```
```live:eg/data/incorrect_no_bitcoin_some:output
```

The reason the rule is incorrect is that variables in Rego are _existentially
quantified_. This means that rule bodies and queries express FOR ANY and not FOR
ALL. To express FOR ALL in Rego complement the logic in the rule body (e.g.,
`!=` becomes `==`) and then complement the check using negation (e.g.,
`no_bitcoin_miners` becomes `not any_bitcoin_miners`).

Alternatively, we can implement the same kind of logic inside a single rule
using [Comprehensions](#comprehensions).

Expand All @@ -1126,26 +1150,14 @@ no_bitcoin_miners_using_comprehension {
}
```

By importing the future keyword "every", you get another option to express universal
quantification:

```live:eg/data/every_alternative:module:read_only
import future.keywords.every
no_bitcoin_miners_using_every {
every app in apps {
app.name != "bitcoin-miner"
}
}
```

{{< info >}}
Whether you use negation, comprehensions, or `every` to express FOR ALL is up to you.
The `every` keyword should lend itself nicely to a rule formulation that closely
follows how requirements are stated, and thus enhances your policy's readability.

The comprehension version is more concise than the negation variant, and does not
require a helper rule while the negation version is more verbose but a bit simpler and allows for more complex ORs.
require a helper rule while the negation version is more verbose but a bit simpler
and allows for more complex ORs.
{{< /info >}}

## Modules
Expand Down Expand Up @@ -1362,18 +1374,18 @@ scope of the body evaluation:
```live:eg/every1:module:merge_down
import future.keywords.every
p {
array_domain {
every i, x in [1, 2, 3] { x-i == 1 } # array domain
}
q {
object_domain {
every k, v in {"foo": "bar", "fox": "baz" } { # object domain
startswith(k, "f")
startswith(v, "b")
}
}
r {
set_domain {
every x in {1, 2, 3} { x != 4 } # set domain
}
```
Expand All @@ -1387,25 +1399,25 @@ construct using a helper rule:
import future.keywords.every
xs := [2, 2, 4, 8]
p(x) := x > 1
larger_than_one(x) := x > 1
r {
every x in xs { p(x) }
rule_every {
every x in xs { larger_than_one(x) }
}
s {
not_less_or_equal_one {
not lte_one
}
lte_one {
some x in xs
not p(x)
not larger_than_one(x)
}
```
```live:eg/every2:output
```

Negating `every` is forbidden. If you desire to express `not every x in xs { p(x) }`,
Negating `every` is forbidden. If you desire to express `not every x in xs { p(x) }`
please use `some x in xs; not p(x)` instead.

## With Keyword
Expand Down

0 comments on commit 1ca73a0

Please sign in to comment.