Risk-adjustment techniques for generating software code data
A risk-sensitive code generation computing system receives query data indicating a requested code section. A code attribute evaluation module determines attributes that indicate characteristics of the query data or the requested code section. Based on the attributes, the code attribute evaluation module calculates risk level data for the requested code section. A code generation module selects, based on the risk level data, a set of training resources. Based on the selected training resources, the risk-sensitive code generation computing system modifies training of a trained neural network model. The modified trained neural network model generates a risk-sensitive code data object based on the requested code section and the risk level data. The risk-sensitive code generation computing system provides at least a portion of the risk-sensitive code data object to an additional computing system.
1 . A system comprising:
a processor, and
a non-transitory computer-readable storage device storing instructions that are executable by the processor to:
receive, from a user computing system, query data indicating a requested code subsection;
determine, via a code attribute evaluation module, a set of attributes indicating one or more characteristics of the query data or the requested code subsection;
calculate, based on the set of attributes and via the code attribute evaluation module, a set of risk levels for the requested code subsection;
select, based on the set of risk levels and via a code generation module, a subset of training resources from a group of multiple code-generation training resources;
based on the selected subset of training resources, modify training of a trained text generation model that is configured to generate text data indicating a risk-sensitive code subsection based on the requested code subsection;
receive, from the modified trained text generation model, a risk-sensitive code data object that includes the text data indicating the risk-sensitive code subsection; and
provide the risk-sensitive code data object to the user computing system.
2 . The system of claim 1 , wherein calculating the set of risk levels further comprises:
accessing one or more risk event data collections;
determining a correlation between a particular attribute from the set of attributes and particular risk data included in the risk event data collections; and
determining, based on the correlation, at least one risk level value indicating a relative risk associated with the particular attribute.
3 . The system of claim 1 , the instructions further executable to:
access historical data associated with a previous code version associated with the requested code subsection;
identify risk event data correlated with the previous code version;
calculate a causation probability indicating a likelihood that the previous code version was modified in response to an event described by the risk event data; and
modify the set of risk levels based on the causation probability.
4 . The system of claim 3 , wherein the historical data includes one or more of: the previous code version, change request data associated with the previous code version, or documentation data associated with the previous code version.
5 . The system of claim 1 , the instructions further executable to:
determine, based on the set of risk levels, a recommended review value associated with the risk-sensitive code data object; and
responsive to determining that the recommended review value exceeds a review threshold value, provide the risk-sensitive code data object to an additional user computing system.
6 . The system of claim 5 , the instructions further executable to, responsive to determining that the recommended review value exceeds a testing threshold level, generate one or more testing data objects associated with the risk-sensitive code data object.
7 . The system of claim 1 , wherein the risk-sensitive code data object includes additional text data indicating comments, wherein determining the comments comprises:
determining, based on the set of attributes and by the code generation module, a comment detail level associated with one or more code segments included in the requested code subsection; and
generating, based on the comment detail level and by the code generation module, the additional text data indicating comments.
8 . The system of claim 1 , wherein the set of attributes includes at least one attribute associated with one or more of:
a security vulnerability associated with the requested code subsection;
an additional code subsection with which the requested code subsection is configured to interoperate;
an organization policy associated with the requested code subsection; or
an experience level of a user associated with the user computing system.
9 . A method including operations executed by a processor, the operations comprising:
receiving, from an additional computing system, query data indicating a requested code section;
determining a set of attributes indicating one or more characteristics of the query data or the requested code section;
calculating, based on the set of attributes, a set of risk levels for the requested code section;
selecting, based on the set of risk levels, a subset of training resources from multiple code-generation training resources;
based on the selected subset of training resources, modifying training of a trained neural network model that is configured to generate data indicating a risk-sensitive code section that is based on the requested code section;
receiving, from the modified trained neural network model, a risk-sensitive code data object that includes the data indicating the risk-sensitive code section; and
providing the risk-sensitive code data object to the additional computing system.
10 . The method of claim 9 , wherein calculating the set of risk levels further comprises:
accessing one or more risk event data collections;
determining a correlation between a particular attribute from the set of attributes and particular risk data included in the risk event data collections; and
determining, based on the correlation, at least one risk level value indicating a relative risk associated with the particular attribute.
11 . The method of claim 9 , the operations further comprising:
accessing historical data associated with a previous code version associated with the requested code section;
identifying risk event data correlated with the previous code version;
calculating a causation probability indicating a likelihood that the previous code version was modified in response to an event described by the risk event data; and
modifying the set of risk levels based on the causation probability.
12 . The method of claim 9 , the operations further comprising:
determining, based on the set of risk levels, a recommended review value associated with the risk-sensitive code data object; and
responsive to determining that the recommended review value exceeds a review threshold value, providing the risk-sensitive code data object to an additional user computing system.
13 . The method of claim 12 , the operations further comprising, responsive to determining that the recommended review value exceeds a testing threshold level, generate one or more testing data objects associated with the risk-sensitive code data object.
14 . The method of claim 9 , wherein the risk-sensitive code data object includes text data indicating comments, wherein determining the comments comprises:
determining, based on the set of attributes, a comment detail level associated with one or more code segments included in the requested code section; and
generating, based on the comment detail level, the text data indicating the comments.
15 . The method of claim 9 , wherein the set of attributes includes at least one attribute associated with one or more of:
a security vulnerability associated with the requested code section;
an additional code section with which the requested code section is configured to interoperate;
an organization policy associated with the requested code section; or
an experience level of a user associated with the additional computing system.
16 . A non-transitory computer-readable medium embodying program code, wherein, when executed by a processor, the program code causes the processor to perform operations comprising:
receiving, from an additional computing system, query data indicating a requested code section;
determining a set of attributes indicating one or more characteristics of the query data or the requested code section;
calculating, based on the set of attributes, a set of risk levels for the requested code section;
selecting, based on the set of risk levels, a subset of training resources from multiple code-generation training resources;
based on the selected subset of training resources, modifying training of a trained neural network model that is configured to generate data indicating a risk-sensitive code section that is based on the requested code section;
receiving, from the modified trained neural network model, a risk-sensitive code data object that includes the data indicating the risk-sensitive code section; and
providing the risk-sensitive code data object to the additional computing system.
17 . The non-transitory computer-readable medium of claim 16 , wherein calculating the set of risk levels further comprises:
accessing one or more risk event data collections;
determining a correlation between a particular attribute from the set of attributes and particular risk data included in the risk event data collections; and
determining, based on the correlation, at least one risk level value indicating a relative risk associated with the particular attribute.
18 . The non-transitory computer-readable medium of claim 16 , the operations further comprising:
accessing historical data associated with a previous code version associated with the requested code section;
identifying risk event data correlated with the previous code version;
calculating a causation probability indicating a likelihood that the previous code version was modified in response to an event described by the risk event data; and
modifying the set of risk levels based on the causation probability.
19 . The non-transitory computer-readable medium of claim 16 , the operations further comprising:
determining, based on the set of risk levels, a recommended review value associated with the risk-sensitive code data object; and
responsive to determining that the recommended review value exceeds a review threshold value, providing the risk-sensitive code data object to an additional user computing system.
20 . The non-transitory computer-readable medium of claim 16 , wherein the set of attributes includes at least one attribute associated with one or more of:
a security vulnerability associated with the requested code section;
an additional code section with which the requested code section is configured to interoperate;
an organization policy associated with the requested code section; or
an experience level of a user associated with the additional computing system.