Systems and methods for generating textual instructions for manufacturers from hybrid textual and image data
A system for generating textual instructions for manufacturers from hybrid textual and image data includes a manufacturing instruction generator that may generate a language processing module from a first training set including at least a training annotated file describing at least a first product to manufacture, the at least an annotated file containing one or more textual data, and at least an instruction set containing one or more manufacturing instructions to manufacture the at least a first product. Manufacturing instruction generator may use the language processing to generate textual instructions for manufacturers from at least an annotated file and may initiate manufacture using the generated manufacturing instructions.
1 . A method of generating textual instructions for manufacturers from hybrid textual and image data, the method performed by at least a computing device and comprising:
generating, from a dataset including a plurality of textual data extracted from annotated files describing products to manufacture and a plurality of correlated textual instructions for manufacturers, at least a language model, wherein the at least a language model receives textual data as inputs and produces textual instructions for manufacturers as outputs;
receiving at least a geometric model and at least an annotated file describing at least a product to manufacture;
extracting, by the manufacturing instruction generator, digital character data from the at least an annotated file;
calculating at least an extraction quality score of the digital character data as a function of an extraction classifier;
determining at least an interrogator output from the at least a geometric model; and
generating at least a textual instruction from the extracted words using the at least a language model and the at least an interrogator output.
2 . The method of claim 1 , wherein calculating at least an extraction quality comprises training the extraction classifier using extraction training data comprises sets of digital character data correlated to extraction quality scores.
3 . The method of claim 1 , wherein the method further comprises:
comparing the extraction quality score to an extraction quality score threshold;
rotating an orientation of that at least an annotated file if the extraction quality score is below the extraction quality score threshold; and
extracting, after rotation, by the manufacturing instruction generator, new digital character data from the at least an annotated file.
4 . The method of claim 3 , wherein rotating the orientation of the at least an annotated file comprises rotating the orientation of the at least an annotated file by 90 degrees.
5 . The method of claim 1 , wherein the extraction classifier comprises a decision tree model.
6 . The method of claim 2 , wherein the extraction training data further comprises extraction metrics correlated to extraction quality scores.
7 . The method of claim 6 , wherein the extraction metrics comprise a percent of words in a set of the sets of digital character data that are in a desired language.
8 . The method of claim 6 , wherein the extraction metrics comprise a percent of the total characters in a set of the sets of digital character data that are whitespace.
9 . The method of claim 1 , further comprising presenting, using a user interface, an extraction list to a user.
10 . The method of claim 9 , further comprising:
receiving a user input, wherein the user input comprises a selection of an item in the extraction list;
displaying, using the user interface, the at least an annotated file; and
manipulating the user interface to display a location associated with the selection of the item in the at least an annotated file.
11 . A system for generating textual instructions for manufacturers from hybrid textual and image data, the system comprising at least a computing device designed and configured to:
generate, from a dataset including a plurality of textual data extracted from annotated files describing products to manufacture and a plurality of correlated textual instructions for manufacturers, at least a language model, wherein the at least a language model receives textual data as inputs and produces textual instructions for manufacturers as outputs;
receive at least a geometric model and at least an annotated file describing at least a product to manufacture;
extract digital character data from the at least an annotated file;
calculate at least an extraction quality score of the digital character data as a function of an extraction classifier;
determine at least an interrogator output from the at least a geometric model; and
generate at least a textual instruction from the extracted words using the at least a language model and the at least an interrogator output.
12 . The system of claim 11 , wherein calculating at least an extraction quality comprises training the extraction classifier using extraction training data comprises sets of digital character data correlated to extraction quality scores.
13 . The system of claim 11 , wherein the computing device is further configured to:
compare the extraction quality score to an extraction quality score threshold;
rotate an orientation of that at least an annotated file if the extraction quality score is below the extraction quality score threshold; and
extract, after rotation, by the manufacturing instruction generator, new digital character data from the at least an annotated file.
14 . The system of claim 13 , wherein rotating the orientation of the at least an annotated file comprises rotating the orientation of the at least an annotated file by 90 degrees.
15 . The system of claim 11 , wherein the extraction classifier comprises a decision tree model.
16 . The system of claim 12 , wherein the extraction training data further comprises extraction metrics correlated to extraction quality scores.
17 . The system of claim 16 , wherein the extraction metrics comprise a percent of words in a set of the sets of digital character data that are in a desired language.
18 . The system of claim 16 , wherein the extraction metrics comprise a percent of the total characters in a set of the sets of digital character data that are whitespace.
19 . The system of claim 11 , wherein the computing device is further configured to present, using a user interface, an extraction list to a user.
20 . The system of claim 19 , wherein the computing device is further configured to:
receive a user input, wherein the user input comprises a selection of an item in the extraction list;
display, using the user interface, the at least an annotated file; and
manipulate the user interface to display a location associated with the selection of the item in the at least an annotated file.